exclusion as a core competition concern - digital commons

American University Washington College of Law American University Washington College of Law

Digital Commons @ American University Washington College of Digital Commons @ American University Washington College of

Law Law

Articles in Law Reviews & Other Academic Journals Scholarship & Research

2013

Exclusion as a Core Competition Concern Exclusion as a Core Competition Concern

Jonathan Baker

Follow this and additional works at: https://digitalcommons.wcl.american.edu/facsch_lawrev

Part of the Antitrust and Trade Regulation Commons, and the Law and Economics Commons

https://digitalcommons.wcl.american.edu/

https://digitalcommons.wcl.american.edu/

https://digitalcommons.wcl.american.edu/facsch_lawrev

https://digitalcommons.wcl.american.edu/facsch_lawrev

https://digitalcommons.wcl.american.edu/facsch

https://digitalcommons.wcl.american.edu/facsch_lawrev?utm_source=digitalcommons.wcl.american.edu%2Ffacsch_lawrev%2F1130&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/911?utm_source=digitalcommons.wcl.american.edu%2Ffacsch_lawrev%2F1130&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/612?utm_source=digitalcommons.wcl.american.edu%2Ffacsch_lawrev%2F1130&utm_medium=PDF&utm_campaign=PDFCoverPages

EXCLUSION AS A CORE COMPETITION CONCERN

JONATHAN B. BAKER*

Exclusionary conduct1 is commonly relegated to the periphery in contem-porary antitrust discourse, while price fixing, market division, and other formsof collusion are placed at the core of competition policy. When the term “hardcore” is applied to an antitrust violation,2 or the “supreme evil” of antitrust isidentified,3 the reference is invariably to cartels.4 At the same time, antitrust is“more cautious” in condemning exclusion than collusion.5

* Professor of Law, American University Washington College of Law. This paper revises andextends keynote remarks delivered to the Twenty-second Annual Workshop of the CompetitionLaw and Policy Institute of New Zealand (CLPINZ). The author is especially grateful to AndyGavil and also indebted to Svend Albaek, Rick Brunell, Peter Carstensen, Pat DeGraba, AaronEdlin, Harry First, Scott Hemphill, Heather Hughes, Al Klevorick, Prasad Krishnamurthy, BobLande, James May, Doug Melamed, Doug Richards, Steve Salop, David Snyder, Josh Soven,Peter Taylor, John Woodbury, Josh Wright, an anonymous referee, and participants in the facultybusiness law workshop at American University, the law and economics workshop at BerkeleyLaw School, the CLPINZ workshop, and the Loyola Antitrust Colloquium.

1 The terms “exclusion” and “foreclosure,” which will be used interchangeably, encompassboth the complete foreclosure of rivals or potential entrants and conduct that disadvantages rivalswithout necessarily inducing them to exit. Exclusion is anticompetitive if the excluding firms useit to obtain or maintain market power, as by raising price or keeping a supracompetitive pricefrom declining.

2 E.g., ORGANIZATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT, HARD CORE CAR-

TELS 58 (2000), available at http://www.oecd.org/dataoecd/39/63/2752129.pdf (“[H]ard core car-tels are the most egregious violations of competition law . . . .”). “Hard core cartels” are collusivearrangements lacking an efficiency justification. See id. at 6. In Europe, the term “hard core” isalso applied to a class of prohibited vertical restraints. See Commission Regulation (EU) No.303/2010, 2010 O.J. (L 102) 1, 4–5 (stating block exemption for vertical agreements not appliedto supply or distribution agreements that contain a “hardcore” restriction such as vertical pricefixing or territorial or customer sales restrictions).

3 Verizon Commc’ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 408 (2004)(noting that collusion is the “supreme evil” of antitrust).

4 Accord U.S. DEP’T OF JUSTICE, ANTITRUST ENFORCEMENT AND THE CONSUMER 3, availableat http://www.justice.gov/atr/public/div_stats/antitrust-enfor-consumer.pdf (“The worst antitrustoffenses are cartel violations[.]”).

5 HERBERT HOVENKAMP, THE ANTITRUST ENTERPRISE 24 (2005); see Leegin CreativeLeather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 888 (2007) (“Our recent cases formulate anti-trust principles in accordance with the appreciated differences in economic effect between verti-cal and horizontal agreements[.]”); Copperweld Corp. v. Independence Tube Corp., 467 U.S.

527

78 Antitrust Law Journal No. 3 (2013). Copyright 2013 American Bar Association. Reproduced by permission. All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or downloaded or stored in an electronic database or retrieval system without the express written consent of the American Bar Association.

528 ANTITRUST LAW JOURNAL [Vol. 78

Antitrust commentators associated with the Chicago School6 have long ex-pressed deep skepticism about exclusion as an antitrust theory, particularly asapplied to dominant firm conduct.7 Mainstream and progressive commentatorsalso call collusion the central antitrust problem,8 although post-Chicago com-mentators tend more than most to take exclusionary conduct seriously.9 More-over, the antitrust enforcement agencies routinely emphasize collusion overexclusion in articulating their enforcement priorities.10 These rhetorical dis-

752, 768 (1984) (“Concerted activity subject to § 1 is judged more sternly than unilateral activityunder § 2.”).

6 The three major eras of antitrust interpretation—classical (1890 to the 1940s), structural(1940s through the 1970s), and Chicago School (since the late 1970s)—and emerging post-Chi-cago approaches are surveyed in Jonathan B. Baker, A Preface to Post-Chicago Antitrust, inPOST-CHICAGO DEVELOPMENTS IN ANTITRUST LAW 60, 60–67 (Antonio Cucinotta, Roger Vanden Bergh & Roberto Pardolesi eds., 2002).

7 In Judge Robert Bork’s view, in his influential book The Antitrust Paradox, courts shouldalmost never credit the possibility that a firm could exclude rivals by refusing to deal with sup-pliers or distributors also or otherwise force rivals to bear higher distribution costs. ROBERT H.BORK, THE ANTITRUST PARADOX 156, 346 (1978). However, Bork did identify one case in whichhe believed that unilateral conduct by a dominant firm had properly been condemned as exclu-sionary. See id. at 344–46 (citing Lorain Journal Co. v. United States, 342 U.S. 143 (1951)).Judge Richard Posner has similarly described anticompetitive exclusion as “rare,” RICHARD POS-

NER, ANTITRUST LAW 194 (2d ed. 2001), though he is not as skeptical about exclusion as otherChicago School commentators, see id. at 194 & n. 2.

8 For example, Professor Herbert Hovenkamp, author of the leading antitrust treatise, recentlydescribed price fixing as “kind of the first-degree murder of antitrust violations,” Thomas Catan,Critics of E-Books Lawsuit Miss the Mark, Experts Say, WALL ST. J., Apr. 23, 2012, at B1, andProfessor Robert Lande, a Director of the pro-enforcement American Antitrust Institute, recentlycalled collusion among rivals “the essence of the most evil thing we have in antitrust,” SaraForden, U.S. Sues Apple for eBook Pricing as Three Firms Settle, BLOOMBERG (Apr. 17, 2012),http://www.bloomberg.com/news/2012-04-17/u-s-sues-apple-for-ebook-pricing-as-three-firms-settle.html.

9 Professor Steven Salop, a leading post-Chicago antitrust commentator, has long emphasizedthe importance of antitrust’s concern with exclusion. See, e.g., Steven C. Salop, Economic Analy-sis of Exclusionary Vertical Conduct: Where Chicago Has Overshot the Mark, in HOW THE

CHICAGO SCHOOL OVERSHOT THE MARK: THE EFFECT OF CONSERVATIVE ECONOMIC ANALYSIS

ON U.S. ANTITRUST 141, 141–44 (Robert Pitofsky ed., 2008); Thomas G. Krattenmaker &Steven C. Salop, Anticompetitive Exclusion: Raising Rivals’ Costs to Achieve Power Over Price,96 YALE L.J. 209, 213 (1986). See generally Jonathan B. Baker, Professor of Law, Am. Univ.Wash. Coll. of Law, Remarks at the Presentation of the American Antitrust Institute AntitrustAchievement Award to Steven C. Salop (June 24, 2010), http://www.antitrustinstitute.org/~antitrust/sites/default/files/Baker%20Salop%20Comments_062820101005.pdf. Exclusionaryconduct has been the source of the most significant divide between Chicago School and post-Chicago commentators.

10 A recent Assistant Attorney General for Antitrust identified cartel enforcement as hisagency’s “top priority,” well ahead of “single firm conduct” (which often involves exclusion by adominant firm). See Thomas O. Barnett, Deputy Assistant Att’y Gen. for Antitrust, U.S. Dep’t ofJustice, Antitrust Enforcement Priorities: A Year in Review (Nov. 19, 2004), available at http://www.justice.gov/atr/public/speeches/206455.htm. Similarly, it is “uncontroversial,” according toa former Chairman of the FTC, that non-merger antitrust enforcement should focus on “horizon-tal activities” (which are often collusive). Timothy J. Muris, Chairman, Fed. Trade Comm’n,Antitrust Enforcement at the Federal Trade Commission: In a Word—Continuity (Aug. 7, 2001),http://www.ftc.gov/speeches/muris/murisaba.shtm. By “horizontal activities,” Muris intended torefer primarily to collusive conduct; he would have used the term “single firm” to discuss most

2013] EXCLUSION AS A CORE COMPETITION CONCERN 529

tinctions may be framed in terms of traditional doctrinal distinctions betweenconcerted and unilateral conduct, and between horizontal and vertical con-duct, but, as Part I of this article explains, they are better understood in termsof the related but not identical economic distinction between collusive andexclusionary conduct.

Exclusion is routinely described as having a lesser priority than collusioneven though exclusion is well established as a serious competitive problem inboth antitrust law and industrial organization economics. Part II of this articlesurveys the breadth of practices that courts have considered exclusionary,shows that exclusion has not been downplayed in court decisions, and ex-plains that the emerging doctrinal rules governing exclusion and collusionplace the two types of competitive problems on similar footings. In formalstructure, antitrust rules are not tougher on collusion. Rather, the rules aretough on conduct with no plausible efficiency justification, i.e., what is com-monly termed “naked” collusion or what will be referred to here as “plain”exclusion. Part III demonstrates that collusion and exclusion are also closelyrelated as a matter of economics—so much so as to make the economic rea-sons for concern about anticompetitive collusion equally reasons for concernabout anticompetitive exclusion. If anything, as this Part further explains, an-ticompetitive exclusion may be the more important problem because of theparticular threat exclusion poses to economic growth.

Notwithstanding the broad parallels in the economic analysis of exclusionand collusion, the two types of anticompetitive conduct arise through differenteconomic mechanisms. Just as colluding firms must find a way to solve “car-tel problems” (reaching consensus on terms of coordination, deterring cheat-ing on those terms, and preventing new competition), excluding firms mustfind a way to solve “exclusion problems” (identifying an exclusionarymethod, excluding sufficient rivals to harm competition, and ensuring that theexclusionary conduct is profitable for each excluding firm). Despite these dif-ferences, as Part IV demonstrates, the doctrinal rules identified in Part II trun-cate the comprehensive reasonableness analysis of exclusionary conduct inways analogous to the structured reasonableness rules governing collusiveconduct by making it unnecessary to show how or whether defendants solveall the relevant exclusion problems or cartel problems. Again, therefore, the

exclusionary behavior. E-mail from Timothy J. Muris, Former Chairman, Fed. Trade Comm’n,to Jonathan Baker, Professor of Law, Am. Univ. Wash. Coll. of Law (Dec. 10, 2011, 7:37 PMEST) (on file with author). Even enforcers “not so aligned with the Chicago School may ap-proach exclusionary claims more cautiously than collusion claims.” John Woodbury, PaperTrail: Working Papers and Recent Scholarship, ANTITRUST SOURCE, Apr. 2012, http://www.americanbar.org/content/dam/aba/publishing/antitrust_source/apr12_papertrail_4_26f.authcheckdam.pdf (referencing Justice Department materials from the current administration).


formal structure of antitrust rules does not downplay anticompetitiveexclusion.

The rhetorical consensus is so powerful that claims of priority for collusionover exclusion are typically stated without explicit justification. They never-theless appear to be grounded primarily in two commonly accepted andclosely related suppositions, evaluated critically in Part V along with otherpurported justifications for downplaying exclusion. The first supposition isthat it is more difficult for courts and enforcers to identify anticompetitiveexclusionary conduct than to identify harmful collusive conduct, because con-duct that looks exclusionary commonly also promotes competition by enhanc-ing efficiency.11 The second is that exclusionary conduct often benefitsconsumers in the short run, and that as a result, overly aggressive enforcementagainst exclusionary conduct risks chilling such procompetitive practices asprice cutting and new product introductions. Together, these premises, if ac-cepted, would imply that mistakes in enforcement and adjudication againstanticompetitive exclusion pose greater threats than mistakes in enforcementand adjudication against anticompetitive collusion, and, consequently, wouldjustify downplaying exclusionary conduct in antitrust enforcement.

Justice Scalia’s opinion for the Supreme Court in Verizon CommunicationsInc. v. Law Offices of Curtis V. Trinko also suggests that mistakes in enforce-ment and adjudication are more frequent and more troublesome in exclusioncases, but grounds that view in different and more controversial arguments.12

The opinion’s sweeping rhetoric—all dicta13—minimizes the competitive

11 See, e.g., U.S. DEP’T OF JUSTICE, COMPETITION AND MONOPOLY: SINGLE-FIRM CONDUCT

UNDER SECTION 2 OF THE SHERMAN ACT 12–13 (2008), available at http://www.justice.gov/atr/public/reports/236681.pdf, withdrawn, Press Release, U.S. Dep’t of Justice, Justice DepartmentWithdraws Report on Antitrust Monopoly Law (May 11, 2009), http://www.usdoj.gov/atr/public/press_releases/2009/245710.pdf [hereinafter DOJ SECTION 2 REPORT (WITHDRAWN)]. Thesethemes were also emphasized by speakers at a conference discussion of a draft of this article,including among those sympathetic to a robust antitrust concern with exclusionary conduct.

12 See generally Verizon Commc’ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S.398 (2004). The central claim in the case involved exclusionary conduct: a class of local tele-phone service customers alleged that Verizon, an incumbent local exchange carrier, had pro-tected its monopoly prices from erosion by denying interconnection services to entrants seekingto offer competing local telephone service. Id. at 402–03 (noting that Verizon was obligated toprovide new entrants with interconnection services under the Telecommunications Act of 1996).The Court held that Verizon’s unilateral refusal to assist its rivals did not state a claim under theSherman Act. Id. at 415–16.

13 Trinko is best read as precluding monopolization liability in a setting in which a separatestatutory scheme provides for extensive regulation aimed at promoting competition. (The statuteincorporated specific mechanisms for promoting competition by requiring incumbent monopo-lists to deal with entrants.) See id. at 402–03. If the regulatory scheme is sufficiently extensiveand effective, Trinko holds, antitrust enforcement may be displaced. See id. at 413 (“[T]he [regu-latory] regime was an effective steward of the antitrust function.”); see also Nobody v. ClearChannel Commc’ns, Inc., 311 F. Supp. 2d 1048, 1112–14 (D. Colo. 2004) (limiting Trinko toregulated industry settings). But see John Doe 1 v. Abbott Labs., 571 F.3d 930, 934 (9th Cir.


concern arising from a monopolist’s unilateral exclusionary acts by implicitlydescribing the consequences of judicial mistakes from Sherman Act Section 2enforcement in asymmetric terms.14 The Trinko opinion can be understood toclaim that false negatives (false acquittals) are not troublesome because mo-nopolies are temporary, hence self-correcting;15 that false positives (false con-victions) are troublesome because monopolies foster economic growth;16 andthat courts cannot practically craft relief to avoid ongoing judicial supervision,at least with respect to the violation alleged in the case.17 Consistent with itsskeptical view of antitrust enforcement against exclusionary conduct, Trinkodeclares that collusion is the “supreme evil” of antitrust18 and rhetorically cab-ins-in an earlier pro-plaintiff monopolization decision, Aspen Skiing Co. v.Aspen Highlands Skiing Corp.,19 by describing it as “at or near the outerboundary of § 2 liability.”20

Whether the claimed policy justifications for downplaying exclusion aregrounded in common ideas about the difficulty distinguishing procompetitiveconduct from anticompetitive exclusion or in the more controversial argu-ments made in Trinko, they do not stand up to analysis. The antitrust enforce-ment agencies do challenge collusion more frequently than exclusion. Thisobservation could be explained through a theory sympathetic to assigning ex-clusion a lower priority than collusion, as consistent with the dual supposi-tions that it is more difficult to rule out efficiencies and avoid erroneousfindings of liability in exclusionary conduct cases than in collusive conductcases. But, as Part V explains, the relatively low frequency of enforcementagainst anticompetitive exclusion instead probably reflects an enforcement

2009) (considering Trinko outside of the regulated industries context). More recently, the Courtagain displaced an antitrust court in favor of awarding exclusive jurisdiction over competitionenforcement to an industry regulator. Credit Suisse Sec. (USA) LLC v. Billing, 551 U.S. 264(2007) (expanding the implied antitrust immunity conferred by regulation under the securitieslaws). See generally Howard A. Shelanski, The Case for Rebalancing Antitrust and Regulation,109 MICH. L. REV. 683 (2011) (discussing the reasoning and potential consequences of Trinkoand Credit Suisse).

14 See Andrew I. Gavil, Exclusionary Distribution Strategies by Dominant Firms: Striking aBetter Balance, 72 ANTITRUST L.J. 3, 42–51 (2004) (offering a detailed exposition and critiqueof the rhetoric of the majority opinion in Trinko).

15 See Trinko, 540 U.S. at 407.16 The opinion argues in particular that the prospect of monopoly induces risk-taking and

innovation. See id.17 See id. at 414–15.18 Id. at 408.19 472 U.S. 585 (1983).20 Trinko, 540 U.S. at 409; accord Pacific Bell Tel. Co. v. linkLine Commc’ns, Inc., 555 U.S.

438, 448 (2009) (dictum) (“[Aspen suggests that] a firm’s unilateral refusal to deal with its rivalscan give rise to antitrust liability [in] limited circumstances[.]”). Appeals courts have noted thenarrow reading that Trinko and linkLine give to Aspen. See, e.g., Broadcom Corp. v. QualcommInc., 501 F.3d 297, 316 (3d Cir. 2007); MetroNet Servs. Corp. v. Qwest Corp., 383 F.3d 1124,1131–34 (9th Cir. 2004).


preference for attacking conduct lacking a legitimate justification, combinedwith the greater frequency of naked collusion relative to plain exclusion whenbusinesses operate in the shadow of antitrust enforcement. Part V also criti-cizes other policy arguments that have been offered for assigning lesser prior-ity to exclusion, including one based on empirical studies, another rooted inan analysis of institutional competence, and still others suggested by the Su-preme Court in Trinko.

The troublesome rhetorical consensus placing exclusionary conduct at anti-trust’s periphery, not its core, is not just unwarranted; it is damaging. Themore that exclusion is downplayed rhetorically, the more that its legitimacy asa subject for antitrust enforcement will be undermined,21 so the greater thelikelihood that antitrust rules will eventually change to limit enforcementagainst anticompetitive foreclosure when they should not. Accordingly, an-ticompetitive exclusion, like anticompetitive collusion, must be understood asa core concern of competition policy.

Part VI of this article discusses the implications for antitrust enforcement ofrecognizing exclusion as a core concern of competition policy along with col-lusion. Doing so could lead enforcers to assign a higher priority to attackingexclusion than they do today, particularly challenging conduct foreclosing po-tential entry in markets subject to rapid technological change. In addition, itwould encourage further development of the doctrinal rule governing trun-cated condemnation of exclusionary conduct in the courts, and protect therules governing anticompetitive exclusion against pressure for modificationsthat would limit enforcement.

I. EXCLUSION AS AN ANTITRUST CATEGORY

Exclusion and collusion are neither statutory nor doctrinal categories; theyare economic categories. The Sherman Act distinguishes between concertedconduct (Section 1) and single-firm behavior (Section 2),22 each of whichcould harm competition through exclusion or collusion.23 The doctrinal rulesdeveloped to implement both the Sherman Act and the Clayton Act prohibi-tion on anticompetitive mergers distinguish between horizontal and verticalagreements, each of which again could harm competition through exclusion or

21 Cf. Jonathan B. Baker, Preserving a Political Bargain: The Political Economy of the Non-Interventionist Challenge to Monopolization Enforcement, 76 ANTITRUST L.J. 605, 625–26(2010) (“Had Microsoft come out differently, Trinko might have gone farther to question thelegitimacy of the antitrust bar on monopolization.”).

22 15 U.S.C. §§ 1–2. Sherman Act Section 2 also recognizes conspiracy to monopolize, butthis statutory provision is rarely invoked.

23 A single firm could harm competition collusively if a dominant firm fixes prices or dividesmarkets in cooperation with a fringe rival, for example.


collusion.24 Although these legal categories continue to play a role in modernantitrust analysis, “today’s antitrust lawyers, enforcers and courts focus farmore on the nature of the anticompetitive effects, and in private cases, theantitrust injuries, alleged.”25 For this reason, the antitrust casebook I co-au-thored “separately groups conduct threatening collusive anticompetitive ef-fects—including traditional horizontal agreements, vertical intrabrandagreements and horizontal mergers—and conduct threatening exclusionary ef-fects—including dominant firm behavior, vertical interbrand restraints andvertical mergers.”26 In making this distinction, the casebook adopted the majorstructural division employed by Judge Posner in his antitrust treatise and hisco-authored antitrust casebook.27

Although exclusionary claims are most commonly framed as challenges tovertical agreements or monopolization, antitrust’s traditional doctrinal catego-ries do not perfectly track the distinction between exclusion and collusion.Vertical conduct is not invariably exclusionary. Agreements between manu-facturers and distributors, for example, may harm competition by facilitatingcollusion at either level as well as by excluding entrants into manufacturing ordistribution.28 Nor is horizontal conduct invariably collusive. The category in-cludes, for example, exclusionary group boycotts.29

24 The Sherman Act also distinguishes between exclusionary and exploitative conduct. SeeUnited States v. Grinnell Corp., 384 U.S. 563, 570–71 (1966) (noting that the exercise of marketpower by a firm that obtained it “as a consequence of a superior product, business acumen, orhistoric accident” is not actionable as monopolization). But see Einer Elhauge, Tying, BundledDiscounts, and the Death of the Single Monopoly Profit Theory, 123 HARV. L. REV. 397, 420–26(2009) (arguing that the Supreme Court’s tying jurisprudence is predicated in part on recognitionof the exploitation of monopoly power as a basis for liability). By contrast, in the Europeanapproach to competition policy, a dominant firm can be found to have abused its positionthrough exploitative offenses such as charging higher prices, though such cases are rare. THE ECLAW OF COMPETITION §§ 4.358–361 (Jonathan Faull & Ali Nikpay eds., 2d ed. 2007). It is anopen question whether exploitative conduct could be reached as a violation of FTC Act Section5.

25 ANDREW I. GAVIL, WILLIAM E. KOVACIC & JONATHAN B. BAKER, ANTITRUST LAW IN PER-

SPECTIVE: CASES, CONCEPTS AND PROBLEMS IN COMPETITION POLICY vii (2d ed. 2008) [hereinaf-ter ANTITRUST LAW IN PERSPECTIVE]. Although modern antitrust emphasizes effects, theSherman Act’s agreement requirement means that the statute does not reach every instance inwhich firms harm competition through coordination. See Jonathan B. Baker, Two Sherman ActSection 1 Dilemmas: Parallel Pricing, the Oligopoly Problem, and Contemporary EconomicTheory, 38 ANTITRUST BULL. 143 (1993).

26 ANTITRUST LAW IN PERSPECTIVE, supra note 25, at vii. The doctrinal categories are groupedbased on whether the harm to competition addressed in the leading cases more commonly resultsfrom exclusion or collusion, but all the practices could harm competition either way.

27 See POSNER, supra note 7; RICHARD A. POSNER & FRANK H. EASTERBROOK, ANTITRUST:CASES, ECONOMIC NOTES AND OTHER MATERIALS (2d ed. 1981); cf. BORK, supra note 7, at 134(describing collusive and exclusionary conduct as “two theories of the ways in which competi-tion may be injured that . . . shape and drive the law”).

28 See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 892–94 (2007) (dis-cussing both collusive and exclusionary explanations for resale price maintenance).

29 E.g., Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284 (1985).


The antitrust rules most closely associated with exclusion—those gov-erning the conduct of monopolists and would-be monopolists, and verticalagreements—have long been among the most controversial in U.S. competi-tion policy; the antitrust norms in these categories have aptly been describedas “contested.”30 Over the course of antitrust history, the Supreme Court hasrepeatedly altered its approach to evaluating the legality of vertical non-pricerestraints.31 The modern legal rule nearly inverts the rule applied forty-fiveyears ago.32 The standard used to test vertical agreements concerning price hasbeen even less consistent,33 and remains contested,34 although the case law didnot explicitly associate resale price maintenance with exclusion until re-cently.35 A switch of one vote would have led the Supreme Court to abandonthe longstanding per se prohibition against tying.36 Monopolization standards

30 William E. Kovacic, The Modern Evolution of U.S. Competition Policy EnforcementNorms, 71 ANTITRUST L.J. 377, 410 (2003) (describing as “contested” the antitrust norms gov-erning “abuse of dominance and vertical contractual restraints” between 1961 and 2000).Kovacic sees different patterns in the evolution of norms developed in other areas of antitrust:“progressive contraction (Robinson-Patman matters), progressive expansion (criminal and civilhorizontal restraints), [or] contraction followed by stabilization (mergers) . . . .” Id.

31 See, e.g., Richard A. Posner, The Next Step in the Antitrust Treatment of Restricted Distri-bution: Per Se Legality, 48 U. CHI. L. REV. 6, 6 (1981) (noting that between 1963 and 1976, thelegality of (non-price) distribution restrictions “oscillated from the Rule of Reason to per seillegality and back”).

32 The Court adopted a rule of per se illegality in 1967, in United States v. Arnold, Schwinn &Co., 388 U.S. 365 (1967), but overruled that decision in favor of applying the rule of reason tenyears later, in Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977). Since 1977,vertical non-price restraints have rarely been prohibited, leading one commentator to describe thepractical standard as close to per se legality. See Douglas H. Ginsburg, Vertical Restraints: DeFacto Legality Under the Rule of Reason, 60 ANTITRUST L.J. 67 (1991).

33 Since vertical restraints on price were held illegal per se in Dr. Miles Medical Co. v. JohnD. Park & Sons Co., 220 U.S. 373 (1911), Congress authorized states to allow such agreements(see Miller-Tydings Act, Pub. L. No. 314, 50 Stat. 693 (1937)), then broadened that authority(see McGuire Act, Pub. L. No. 542, 66 Stat. 631 (1952)), and then returned the law to the rule ofper se illegality by repealing its authorization (see Consumer Goods Pricing Act of 1975, Pub. L.No. 94-145, 89 Stat. 801). More recently, the Supreme Court overruled Dr. Miles and adoptedthe rule of reason. Leegin, 551 U.S. at 907.

34 The Leegin decision drew a passionate dissent from four Justices. See id. at 908 (Breyer, J.,dissenting). One month before the oral argument, an FTC Commissioner issued an unusual pub-lic statement detailing her disagreement with the Solicitor General’s pro-defendant brief. SeeLetter from Pamela Jones Harbour, Commissioner, Fed. Trade Comm’n, to the Supreme Court ofthe United States (Feb. 26, 2007), available at http://www.ftc.gov/speeches/harbour/070226verti-calminimumpricefixing.pdf. The continuing controversy over resale price maintenance in thewake of Leegin is discussed in Andrew Gavil, Resale Price Maintenance in the Post-LeeginWorld: A Comparative Look at Recent Developments in the United States and European Union,CPI ANTITRUST J., Summer 2010, Vol. 6, No. 1.

35 See Leegin, 551 U.S. at 892–94 (discussing both collusive and exclusionary explanationsfor resale price maintenance).

36 See Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2 (1984) (5-4 decision). However,courts have found ways to avoid applying the per se prohibition when tying may promote compe-tition. See, e.g., United States v. Microsoft Corp., 253 F.3d 34, 89–95 (D.C. Cir. 2001) (en banc).See generally 1 ABA SECTION OF ANTITRUST LAW, ANTITRUST LAW DEVELOPMENTS 201–04


are also controversial, as is evident from a debate between the federal antitrustenforcement agencies early in the 21st century.37

From a contemporary perspective that recognizes the central role economicconcepts now play in antitrust, the controversies over monopolization and ver-tical restraints standards are best understood as proxy battles over the appro-priate treatment of exclusionary conduct.38 This interpretation draws aneconomic distinction between exclusion and collusion in preference to thestatutory distinction between single firm conduct and agreements, and thedoctrinal distinction between horizontal agreements on the one hand and verti-cal agreements and monopolization on the other.

As antitrust has come to focus on economic distinctions in preference todoctrinal pigeonholes, exclusion has become an important and controversialantitrust category. For this reason, the next Part looks across doctrinal catego-ries in identifying common themes in the judicial approach to exclusionaryconduct.

II. EXCLUSION IN ANTITRUST CASE LAW AND DOCTRINE

The rhetorical consensus downplaying the significance of exclusionary con-duct is surprising because anticompetitive exclusion is treated by antitrust lawas a serious competitive problem. A number of leading U.S. antitrust deci-sions, including recent ones, have been concerned primarily with exclusionaryconduct. Microsoft made it difficult for Netscape to market its browsers tocomputer users in order to protect its Windows operating system monopolyfrom the competition that would be created if software applications could ac-cess any operating system through the browser.39 Standard Oil exploited itsleverage over the railroads to stop the entry of new refiners in order to protectits monopoly in oil refining.40 Before AT&T (the Bell System) was broken up,

(7th ed. 2012) [hereinafter ANTITRUST LAW DEVELOPMENTS]; HERBERT HOVENKAMP, FEDERAL

ANTITRUST POLICY: THE LAW OF COMPETITION AND ITS PRACTICE § 10.7a (4th ed. 2011).37 During the George W. Bush administration, the Justice Department encouraged courts to

adopt a doctrinal approach that would favor defendants, but the Federal Trade Commissionpointedly refused to go along. At the start of the Obama administration, the Justice Departmentwithdrew the previous administration’s proposal. See Baker, supra note 21, at 606–07.

38 See POSNER, supra note 7, at 4 (arguing that the economic theory of monopoly had muchmore to say about collusive practices than exclusionary ones, leading some economists and law-yers identified with the Chicago School (but not Posner) to the view “that there was no economicbasis for concern with the exclusionary practices”). But see Bus. Elecs. Corp v. Sharp Elecs.Corp., 485 U.S. 717, 747–48 (1988) (Stevens, J., dissenting) (explaining the majority opinion asturning, without justification, on treating the distinction between horizontal and vertical agree-ments as more important than the distinction between collusive and exclusionary conduct).

39 Microsoft, 253 F.3d at 64–67.40 United States v. Standard Oil, 221 U.S. 1 (1911). See generally Elizabeth Granitz & Benja-

min Klein, Monopolization by “Raising Rivals Costs”: The Standard Oil Case, 39 J.L. & ECON.1 (1996). Compare George L. Priest, Rethinking the Economic Basis of the Standard Oil Refining


it maintained market power in unregulated markets for specialized telephoneservice and customer premises equipment by discriminating against rivals thatsought to connect with its regulated local telephone service monopoly.41 Visaand MasterCard prevented member banks from issuing American Express andDiscover cards in order to protect their own market power.42

Exclusionary conduct allegations are also central to other antitrust decisionscommonly thought of as alleging collusion. The NCAA threatened two largestate universities with disciplinary action if they did not comply with theNCAA’s arrangement for broadcasting college football games.43 Dentists thatdid not comply with the advertising restrictions promulgated by the CaliforniaDental Association could be censured or expelled.44 The National Society ofProfessional Engineers encouraged state societies to launch disciplinary pro-ceedings against engineers that did not comply with its ethical code, whichincluded the challenged restrictions on competitive bidding restrictions.45 Thepredatory conspiracies alleged (but ultimately not demonstrated) in BrookeGroup and Matsushita were said to have excluded generic cigarettes and aU.S. firm manufacturing televisions, respectively.46 The horizontal (collusive)market division agreement attacked in Topco allowed the firms to preventtheir rivals from selling the cooperative’s private label products.47 The nearlytwo hundred insurance companies indicted for price fixing during the early1940s were also accused of employing boycotts, coercion, and intimidation toprevent competition from firms that were not members of their tradeassociation.48

During the modern era, moreover, the Supreme Court and the appellatecourts have addressed exclusionary conduct without consistently favoring ei-

Monopoly: Dominance Against Competing Cartels, 85 S. CAL. L. REV. 499 (2012) (questioningsome of Granitz & Klein’s argument), with Benjamin Klein, The “Hub-and-Spoke” Conspiracythat Created the Standard Oil Monopoly, 85 S. CAL. L. REV. 459 (2012) (responding to Priest).

41 See United States v. AT&T, 552 F. Supp. 131, 162, 170 (D.D.C. 1982) (entering consentdecree requiring divestitures), aff’d, 714 F.2d 178 (D.C. Cir. 1983).

42 United States v. Visa U.S.A., Inc., 344 F.3d 229, 240 (2d Cir. 2003).43 NCAA v. Bd. of Regents of the Univ. of Okla., 468 U.S 85, 85 (1984).44 Cal. Dental Ass’n v. FTC, 526 U.S. 756 (1959).45 United States v. Nat’l Soc’y of Prof’l Eng’rs, 389 F. Supp. 1193, 1210 (D.D.C. 1974), aff’d

555 F.2d 978 (D.C. Cir. 1977), aff’d 435 U.S. 679 (1978).46 Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993); Matsushita

Elec. Indus. Co., v. Zenith Radio Corp., 475 U.S. 574 (1986). Brooke Group was arguably de-cided incorrectly. See Jonathan B. Baker, Predatory Pricing after Brooke Group: An EconomicPerspective, 62 ANTITRUST L.J. 585, 598 (1994) (“[T]he Court took the case from the jury toaward judgment to the defendant when the record on this key question of fact, construed favora-bly to plaintiff, arguably supported plaintiff’s position.”).

47 United States v. Topco Assocs., 405 U.S. 596 (1972). See generally Peter C. Carstensen &Harry First, Rambling Through Economic Theory: Topco’s Closer Look, in ANTITRUST STORIES

171, 171–204 (Eleanor M. Fox & Daniel A. Crane eds., 2007).48 United States v. South-Eastern Underwriters Ass’n, 322 U.S. 533, 535–36 (1944).


ther defendants or plaintiffs.49 Looking to outcome, reasoning, and tone, deci-sions of the courts of appeals during the first half of 2011 (an arbitrarilychosen recent period),50 as well as notable decisions of the circuit courts andthe Supreme Court from the past three decades,51 do not systematically favoreither side.

To show how seriously the courts take exclusionary conduct, this articleadopts two approaches. Part II.A documents the wide range of exclusionaryconduct that the courts have evaluated. This informal survey shows that an-ticompetitive exclusion has not been downplayed by the courts through limita-tion to a narrow range of practices, contrary to what the rhetoric of enforcersand commentators might suggest.

Part II.B identifies parallels in the formal structure of the emerging doctri-nal rules employed by the courts to identify anticompetitive exclusionary andanticompetitive collusive conduct. In particular, the courts have evolved asimilar approach to the two doctrinal areas: adopting a presumption in eachagainst conduct lacking a plausible efficiency justification. Although these

49 The relative success of plaintiffs and defendants is difficult to interpret because it maydepend on a variety of factors beyond the general attitude of the courts, including whether legalrules are changing, the willingness of firms to engage in questionable conduct that could bechallenged, and the willingness of the parties to a lawsuit to litigate rather than settle. Accord-ingly, even a consistently one-sided pattern of decisions may be a poor indicator of judicialattitudes toward exclusionary conduct.

50 During this period, arguably pro-enforcement exclusion decisions were issued by circuitcourts of appeals in Watson Carpet & Floor Covering, Inc. v. Mohawk Industries, Inc., 648 F.3d452 (6th Cir. 2011), E.I DuPont de Nemours & Co. v. Kolon Industries, Inc., 637 F.3d 435 (4thCir. 2011), and Realcomp II, Ltd. v. FTC, 635 F.3d 815 (6th Cir. 2011), while arguably non-interventionist exclusion decisions were issued in Southeast Missouri Hospital v. C.R. Bard Inc.,642 F.3d 608 (8th Cir. 2011), Brantley v. NBC Universal, Inc., 649 F.3d 1078 (9th Cir. 2011),and Smugglers’ Notch Homeowners Association v. Smugglers’ Notch Management Co., 414 F.App’x 372 (2d Cir. 2011).

51 The Supreme Court exclusion decisions listed below more often take the non-interventionistside, exclusively so since 1993, but the circuit courts do not appear to have interpreted the recentpattern as a mandate to raise the bar to plaintiffs in exclusion cases. Notable decisions from theSupreme Court and appellate courts arguably on the pro-enforcement side of the ledger includeEastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992); Aspen Skiing Co. v.Aspen Highlands Skiing Corp., 472 U.S. 585 (1983); Spirit Airlines, Inc. v. Northwest. Airlines,Inc., 431 F.3d 917 (6th Cir. 2005); United States v. Dentsply International, Inc., 399 F.3d 181(3d Cir. 2005); United States v. Visa, 344 F.3d 229 (2d Cir. 2003); LePage’s Inc. v. 3M, 324 F.3d141 (3d Cir. 2003); United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001) (en banc);and JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775 (7th Cir. 1999). Notable deci-sions arguably on the non-interventionist side include Pacific Bell Telephone Co. v. linkLineCommunications, Inc., 555 U.S. 438 (2009); Verizon Communications Inc. v. Law Offices ofCurtis V. Trinko, LLP, 540 U.S. 398, 408 (2004); Brooke Group Ltd. v. Brown & WilliamsonTobacco Corp., 509 U.S. 209 (1993); Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447 (1993);Matsushita Electric Industrial Co, v. Zenith Radio Corp., 475 U.S. 574 (1986); Jefferson ParishHospital District No. 2 v. Hyde, 466 U.S. 2 (1984); Cascade Health Solutions v. PeaceHealth,515 F.3d 883 (9th Cir. 2008); E&L Consulting, Ltd. v. Doman Industries Ltd., 472 F.3d 23 (2dCir. 2006); United States v. AMR Corp., 335 F.3d 1109 (10th Cir. 2003); and Omega Environ-mental, Inc. v. Gilbarco, Inc., 127 F.3d 1157 (9th Cir. 1997).


presumptions are invoked more frequently in collusion cases than exclusioncases, the difference is attributable to the greater frequency with which anti-trust enforcers and private plaintiffs challenge facially anticompetitive collu-sion, and not to any difference in how tough the rules are when the two typesof anticompetitive conduct are identified. Part II.C explains why those struc-tural parallels will remain as courts clarify open questions regarding theemerging truncated rule governing exclusionary conduct.

A. EXCLUSIONARY PRACTICES IDENTIFIED BY THE COURTS

The courts do not treat anticompetitive exclusion as an unusual event; in-stead they have recognized that exclusionary conduct harming competitioncan take a wide range of forms.52 The anticompetitive possibilities surveyedhave been divided into three broad categories based on the mechanism bywhich exclusion takes place, with an eye toward the economic analysis in PartIII.53 In general, these practices are neither necessarily nor invariably anticom-petitive, as rivals could be excluded without harm to competition, and prac-tices that exclude rivals could help firms lower costs, improve products, orotherwise achieve efficiencies as well as helping them obtain or maintain mar-ket power. The survey is not intended as an inventory of all possible means ofexclusion; rather, it is intended to illustrate the breadth of conduct that couldharm competition through foreclosure.

The practices described in the first two categories exclude rivals by impos-ing a constraint on the latter firms’ conduct, as by raising rivals’ costs or, tosimilar effect, reducing rivals’ access to customers.54 The methods in the firstcategory can be undertaken by the excluding firms acting alone, whetherthrough the unilateral action of a single excluding firm or the joint action of agroup of excluding firms. The methods in the second category require theexcluding firms to coordinate with firms that are not rivals through thepurchase of an exclusionary right. Because coordination is required, the prof-itability of practices in the second category turns in part on factors not rele-vant to the profitability of practices in the other categories, as discussed belowin Part IV.B. In the third category, the excluding firms discourage competition

52 See generally Richard M. Steuer, Foreclosure, in 2 ABA SECTION OF ANTITRUST LAW,ISSUES IN COMPETITION LAW AND POLICY 925, 926–29 (Wayne Dale Collins ed., 2008) (survey-ing case law).

53 For an alternative classification scheme more closely tied to familiar doctrinal categories,see Scott Hemphill & Tim Wu, Parallel Exclusion, 122 YALE L.J. 1182 (2013) (identifying sixmechanisms of foreclosure: simple exclusion, recruiting agents, overbuying an input, tying andbundling, resale price maintenance, and most favored nations provisions).

54 Input foreclosure strategies are commonly thought of as raising rivals’ costs while customerforeclosure strategies are commonly thought of as limiting rivals’ access to the market, but cus-tomer foreclosure strategies can also be understood as another form of raising rivals’ costs be-cause they raise rivals’ costs of distribution.


by altering their rivals’ incentives, in particular by credibly threatening therivals with harm should the latter firms seek to compete aggressively.

1. Constraints Imposed on Rival Conduct

The most obvious anticompetitive exclusionary strategies directly constrainrivals by imposing costs or reducing rivals’ access to customers. A dominantfirm might destroy a fringe rival’s distribution facilities,55 or obtain a monop-oly position through fraudulent acquisition of a patent.56 To similar effect,a vertically integrated dominant firm could redesign its upstream productin order to create an incompatibility for its downstream rival.57 A firmmay also directly exclude its rivals by failing to disclose in advance its patent

55 See Conwood Co. v. U.S. Tobacco Co., 290 F.3d 768, 778 (6th Cir. 2002) (discussing adominant manufacturer of snuff that excluded a fringe rival by destroying its rival’s in-storedisplay racks); Kenneth P. Brevoort & Howard P. Marvel, Successful Monopolization ThroughPredation: The National Cash Register Company, 21 RES. IN L. & ECON. 85 (John B. Kirkwooded., 2004) (discussing a dominant firm maintained its monopoly power in part through espionageand sabotage; federal criminal prosecution settled by consent). Allegedly tortious conduct ac-companied a restriction on access to supply in Watson Carpet & Floor Covering, Inc. v. MohawkIndustries, Inc., 648 F.3d 452, 455 (6th Cir. 2011), which involved “false derogatory accusationsabout [the excluded firm to] potential customers[.]” See also Nat’l Ass’n of Pharm. Mfrs., Inc. v.Ayerst Labs., 850 F.2d 904, 916 (2d Cir. 1988) (noting that dominant firm’s public statementsdisparaging rival’s product would support monopolization claim if “(1) clearly false, (2) clearlymaterial,[ and] (3) clearly likely to induce reasonable reliance”); Int’l Travel Arrangers, Inc. v.W. Airlines, Inc., 623 F.2d 1255, 1262 (8th Cir. 1980) (dominant firm’s exclusionary conductincluded a false, misleading, and deceptive newspaper ad). Managers at one pizza chain wererecently charged with arson after allegedly burning down a rival’s nearby store in order to in-crease sales, though no antitrust violation was apparently charged. Florida Domino’s ManagersCharged with Burning Down Rival Pizza Parlor, FOXNEWS.COM (Oct. 29, 2011), http://www.foxnews.com/us/2011/10/29/florida-dominos-managers-charged-with-burning-down-rival-pizza-parlor/print#ixzz1d1mfXJ3B. Business torts can exclude rivals without harming competition,though, and thus do not necessarily also constitute violations of the antitrust laws.

56 Walker Process Equip., Inc. v. Food Mach. & Chem. Corp., 382 U.S. 172 (1965).57 E.g., C.R. Bard, Inc. v. M3 Sys., Inc., 157 F.3d 1340, 1382–83 (Fed. Cir. 1998). A decep-

tive misrepresentation concerning incompatibility may similarly harm competition if believed.See Joseph Farrell, Janis K. Pappalardo & Howard Shelanski, Economics at the FTC: Mergers,Dominant-Firm Conduct, and Consumer Behavior, 37 REV. INDUS. ORG. 263, 268 (2010)(describing the FTC complaint against Intel resolved by consent settlement as based in part onthis theory).


rights in a technology adopted as an industry standard,58 engaging in shamlitigation,59 or manipulating a regulatory scheme.60

Other methods by which firms can impose constraints that exclude rivalsmay be less direct but equally harmful. A vertically integrated dominant firmcan refuse to sell a key input to rivals,61 or degrade the quality of the input itprovides, as by refusing to sell the highest quality inputs.62 A vertical mergermay threaten anticompetitive exclusion by conferring an incentive for themerged firm unilaterally to foreclose upstream rivals from access to distribu-tion (customer foreclosure) or unilaterally to foreclose downstream rivalsfrom access to a key input (input foreclosure).63 A dominant firm can excludeits rivals by refusing to deal with their suppliers, thereby discouraging thesuppliers from dealing with competing firms.64 A dominant firm that sellscomplementary products can take customers away from an unintegrated rival,thereby reducing the rival’s scale of operations and so raising its costs. Thedominant firm can also accomplish the same end by tying complementaryproducts together,65 offering discounts to buyers purchasing a package of

58 Several related exclusion scenarios are suggested by the case law. All suppose that a stan-dard-setting organization (SSO) selects a particular technology owned by one firm in preferenceto alternative technologies, conditional on a representation by the firm that it does not haveintellectual property covering the standard or that it will abide by a commitment to license on anon-discriminatory basis and charge reasonable royalties if the technology is selected. After thetechnology is incorporated into the standard, and firms adopting the standard make sunk invest-ments to use it (become locked-in), the firm owning the technology acts inconsistently with itscommitment, as by asserting intellectual property rights and charging royalties (see Dell Com-puter Corp., 121 F.T.C. 616 (1996)), charging unreasonably high royalties, or preventing firmsfrom using its intellectual property if they compete with it in the sale of products that incorporatethe standard. See Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297, 316 (3d Cir. 2007).

59 Cal. Motor Transp. Co. v. Trucking Unlimited, 404 U.S. 508 (1972).60 In re K-Dur Antitrust Litig., 686 F.3d 197 (3d Cir. 2012).61 E.g., Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985) (firm con-

trolling three of the four mountains at a leading destination ski resort excluded the companyowning the fourth mountain from participating in a multi-area ski ticket, making it difficult forthe excluded firm to attract customers scheduling week-long ski vacations).

62 Bell Atl. Corp. v. Twombly, 550 U.S. 544 (2007) (major local telephone companies, whichhad different territorial footprints, allegedly acted in concert to evade their statutory obligation tointerconnect with new rivals by making interconnection costly and cumbersome or providing lowquality connections).

63 Applications of Comcast Corporation, General Electric Company, and NBC Universal, Inc.,For Consent to Assign Licenses and Transfer Control of Licensees, Memorandum Opinion andOrder, 26 FCC Rcd. 48 (2011) (noting that Comcast could disadvantage rival video distributorsby denying them access to NBC programming or raising the price, and disadvantage rival pro-gramming suppliers by denying them access to Comcast’s video distribution customers or charg-ing them more), available at transition.fcc.gov/FCC-11-4.pdf.

64 Lorain Journal Co. v. United States, 342 U.S. 143 (1951) (monopolist newspaper refused toaccept ads from firms that advertised on a new radio station).

65 See Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451 (1992) (Kodak alleg-edly tied copier parts to copier service in order to exclude independent service operators). Tyingor bundling may be employed as an exclusionary strategy. See, e.g., Dennis W. Carlton &Michael Waldman, The Strategic Use of Tying to Preserve and Create Market Power in Evolving


products,66 or offering discounts to buyers based on the share of the buyer’stotal input purchases accounted for by the excluding seller.67 Similar exclu-sionary strategies to those set forth above could be employed by a group ofexcluding firms acting collectively to harm a rival, as through an exclusionarygroup boycott,68 parallel exclusionary conduct,69 or pooling weak patents.70

2. Purchase of an Exclusionary Right

The exclusionary strategies in the second category require the involvementof non-excluding firms to raise rivals’ costs, as through vertical agreement. Afirm can foreclose its rivals by contracting with sellers of key inputs, inexpen-sive distribution, or other complementary products or services to raise theprice that rivals must pay for the complement or to deny rivals access to thatproduct entirely.71 A dominant firm may also employ other contracting strate-gies to raise rivals’ costs. It may overbuy a key input to bid up the market

Industries, 33 RAND J. ECON. 194, 209 (2002); John Simpson & Abraham L. Wickelgren, Bun-dled Discounts, Leverage Theory, and Downstream Competition, 9 AM. L. & ECON. REV. 370(2007); Michael D. Whinston, Exclusivity and Tying in U.S. v. Microsoft: What We Know, andDon’t Know, J. ECON. PERSP., Spring 2001, at 63; Michael D. Whinston, Tying, Foreclosure, andExclusion, 80 AM. ECON. REV. 837 (1990). See generally Eliana Garces, An Introduction toTying, Foreclosure, and Exclusion by M.D. Whinston, 8 COMPETITION POL’Y INT’L 145 (2012)(literature survey). Other explanations for tying include price discrimination, which could eitherharm or promote competition, and an effort to achieve efficiencies such as scale or scope econo-mies for sellers or a reduction in transaction costs for buyers. See, e.g., Marius Schwartz &Daniel Vincent, Quantity Forcing and Exclusion: Bundled Discounts and Nonlinear Pricing, in 2ISSUES IN COMPETITION LAW AND POLICY, supra note 52, at 939; David S. Evans & MichaelSalinger, Why Do Firms Bundle and Tie? Evidence from Competitive Markets and Implicationsfor Tying Law, 22 YALE J. ON REG. 37 (2005).

66 Cascade Health Solutions v. PeaceHealth, 515 F.3d 883 (9th Cir. 2008).67 See Farrell et al., supra note 57, at 267 (discussing FTC complaint resolved by consent

settlement based on this theory). Market share discounts could exclude rivals through two eco-nomic mechanisms. First, they may operate like a tax on incremental buyer purchases fromcompetitors. See id. Second, if rivals’ marginal costs increase as their output falls, market sharediscounts (like quantity discounts) could shift sales away from rivals, thereby raising rivals’ costsby denying them economies of scale. See id.

68 E.g., Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284 (1985)(discussing office supply store members of a purchasing cooperative that expelled a rival).

69 See generally Hemphill & Wu, supra note 53.70 See United States v. Singer Mfg. Co., 374 U.S. 174 (1963). The pooling of patents that

would be essential (or otherwise substitutes) if valid could harm competition through exclusionby reducing the likelihood that questionable patents would be reviewed for validity. Poolingcould also benefit competition if used to avoid costly litigation over patent boundaries.

71 Alcoa, the early 20th century aluminum monopolist, entered into contracts with hydroelec-tric power producers that barred the power companies from supplying electricity to other alumi-num manufacturers. See United States v. Aluminum Co. of Am., 148 F.2d 416, 422 (2d Cir.1945) (Alcoa) (describing 1912 government enforcement action); see also United States v.Dentsply Int’l., Inc., 399 F.3d 181 (3d Cir. 2005); United States v. Microsoft Corp., 253 F.3d 34(D.C. Cir. 2001) (exclusionary agreements between Microsoft and Original Equipment Manufac-turers and Internet Access Providers). When the excluded firm is forced to adopt a higher costmethod of distribution, this exclusionary approach is sometimes described as disrupting an opti-mal distribution strategy.


price; this may be worth it if the higher input price forces rivals to exit,72 or ifthe strategy raises competitors’ marginal costs, and thus increases the marketprice by more than the dominant firm’s own average costs rise.73 The domi-nant firm may also exclude rivals by contracting with suppliers to give themonopolist the benefit of any discount the suppliers offer a rival.74

3. Commitment to Tough Competition

In the third category of exclusionary strategies, excluding firms, perhapsespecially dominant firms, scare off competition through commitments thatconvince rivals that aggressive conduct will be met with a strong response.Such a strategy works when the rivals conclude that their best response is tolive and let live—to avoid entry, price cutting, or other competitive movesthat would provoke the giant.75 The leading antitrust example involves preda-tory pricing: a multimarket monopolist may respond aggressively to singlemarket entry, and profit from doing so mainly by discouraging entry in othermarkets, allowing the monopolist to protect its market power there.76 In addi-

72 See, e.g., Weyerhauser Co. v. Ross-Simmons Hardwood Lumber Co., 549 U.S. 312 (2007)(dominant seller of hardwood lumber allegedly protected the market power of its hardwood lum-ber mills by bidding up the price of logs, in order to force a rival mill to exit). Overbuying couldalternatively be viewed as a constraint on rival conduct, and placed in the first category.

73 In general, a firm’s marginal cost is the cost concept relevant to determining its price, whileits average cost is the cost concept relevant to determining its profitability.

74 See, e.g., Complaint at 3–6, United States v. Blue Cross Blue Shield of Mich., 2012 WL4513600 (E.D. Mich. Oct. 1, 2012) (No. 10-CV-14155), available at http://www.justice.gov/atr/cases/f263200/263235.htm; United States v. Delta Dental of R.I., 943 F. Supp. 172, 174–75(D.R.I. 1996). These contractual provisions are termed “most favored nations” or “most favoredcustomer” clauses. They can protect the dominant firm from new competition by making it im-possible for an entrant to obtain key inputs cheaply from suppliers that might have been willingto give the entrant a discount in exchange for a large share of the entrant’s business. Mostfavored customer provisions can also harm competition by facilitating coordination. See gener-ally Jonathan B. Baker, Vertical Restraints with Horizontal Consequences: Competitive Effectsof “Most-Favored-Customer” Clauses, 64 ANTITRUST L.J. 517 (1996).

75 See generally Richard J. Gilbert, Mobility Barriers and the Value of Incumbency, in 1HANDBOOK OF INDUSTRIAL ORGANIZATION 475, 476–530 (Richard Schmalensee & Robert Wil-lig eds., 1989) (excluding firms can make investments that commit them to an aggressive re-sponse to future rivalry, with the consequence that future competition is deterred); Steven C.Salop, Strategic Entry Deterrence, 69 AM. ECON. REV. PAPERS & PROCEEDINGS 335 (1979)(same); see also JOHN SUTTON, SUNK COSTS AND MARKET STRUCTURE (1981) (excluding firmsmay be able to deter entry by raising a new firm’s post-entry marginal costs of production anddistribution, as through investments that have the effect of increasing the sunk investments a newfirm must make on marketing or research and development if it chooses to enter).

76 Spirit Airlines, Inc. v. Nw. Airlines, Inc., 431 F.3d 917, 921–24 (6th Cir. 2005). A predatormay also succeed by convincing lenders or investors no longer to support the prey (“deeppocket” predation), by convincing a prospective entrant that the predator’s costs are too low tomake entry profitable (predation by “cost-signaling”), or by convincing a prospective entrant thatits product will be unattractive to buyers (“test-market” predation). See generally Patrick Bolton,Joseph F. Brodley & Michael H. Riordan, Predatory Pricing: Strategic Theory and Legal Policy,88 GEO. L.J. 2239 (2000); Aaron S. Edlin, Stopping Above-Cost Predatory Pricing, 111 YALE

L.J. 941 (2002). Predatory pricing may also succeed by denying the prey economies of scale


tion, a dominant firm’s contract with suppliers to give the monopolist thebenefit of discounts offered to rivals could be viewed as a commitment by themonopolist to match any price reduction by a rival (as well as falling into theprevious category, purchase of an exclusionary right).

B. PARALLEL LEGAL RULES

The breadth of practices considered exclusionary in the case law suggeststhat the courts take exclusion seriously. The parallel structure of the legalrules governing exclusion and collusion similarly suggests that exclusionaryconduct is not assigned a lower priority in antitrust law, regardless of differ-ences in the relative frequency with which the two types of conduct are chal-lenged. As will be seen, both types of allegations are generally reviewedunder the rule of reason, and in the emerging framework for doing so, courtsemploy analogous methods of truncation based importantly on the absence ofa plausible efficiency justification.77 The parallelism in legal rules is not pri-marily a legacy of antitrust’s historical reliance on doctrinal categories thatencompass both exclusionary and collusive conduct.78 Instead, the truncation

when the prey has fewer captive buyers, Chiara Fumagalli & Massimo Motta, A Simple Theoryof Predation 1–10 (IGIER Working Paper No. 437, 2012), available at http://www.igier.unibocconi.it/files/437.pdf, or by denying the prey demand side scale economies on the other side ofa two-sided platform, Massimo Motta & Helder Vasconcelos, Exclusionary Pricing in a Two-Sided Market (Centre for Econ. Pol’y Research, Discussion Paper No. 9164, 2012). Recent eco-nomic studies provide multiple examples of predatory pricing. See, e.g., Kenneth G. Elzinga &David E. Mills, Predatory Pricing in the Airline Industry: Spirit Airlines v. Northwest Airlines(2005), in THE ANTITRUST REVOLUTION 219 (John E. Kwoka, Jr. & Lawrence J. White eds., 5thed. 2009); Malcolm R. Burns, Predatory Pricing and the Acquisition Cost of Competitors, 94 J.POL. ECON. 266 (1986); David Genesove & Wallace P. Mullin, Predation and Its Rate of Return:The Sugar Industry, 1887–1914, 37 RAND J. ECON. 47 (2006); Fiona Scott Morton, Entry andPredation: British Shipping Cartels 1879–1929, 6 J. ECON. & MGMT. STRATEGY 679, 714(1997); David F. Weiman & Richard C. Levin, Preying for Monopoly? The Case of SouthernBell Telephone Company, 1894–1912, 102 J. POL. ECON. 103 (1994).

77 The terms “truncated” or “structured” refer to a collection of analytical approaches—per serules, quick look rules, presumptions, and burden shifting—that potentially condition liability ona limited factual inquiry rather than requiring courts to engage in a wide-open reasonablenessanalysis. Limiting the factual inquiry is advantageous if it reduces the costs of operating the legalsystem and provides guidance to firms seeking to comply with the antitrust laws and to generalistjudges seeking to enforce those laws—under circumstances in which limiting the evidence con-sidered is unlikely to result in erroneous decisions relative to what a fact-finder would concludefrom a complete factual review. See ANTITRUST LAW IN PERSPECTIVE, supra note 25, at 206; id.at 103–06 (discussing benefits and costs of per se condemnation). But see Abraham L. Wickel-gren, Determining the Optimal Antitrust Standard: How to Think About Per Se Versus Rule ofReason, 85 S. CAL. L. REV. POSTSCRIPT 52, 54 (2012) (noting that, in some settings, more evi-dence may not improve judicial accuracy, and improved accuracy may not improve firm behav-ior). In general, the errors from truncation could go in either direction: truncated rules couldsweep in conduct that should not be condemned, or avoid condemning conduct that should beprohibited. Conduct that avoids condemnation on a quick look can still be reviewed under thecomprehensive rule of reason.

78 Antitrust’s traditional legal categories do not divide perfectly along exclusion vs. collusionlines. See supra Part I. While exclusion cases tend to be framed as vertical agreements or merg-


methods reflect a modern evolution in rule of reason review that applies toboth collusion and exclusion, and shows that courts do not place a higherburden on plaintiffs seeking to demonstrate anticompetitive exclusionaryconduct.

In both the exclusion and the collusion context, the legal rules single out forparticular attention anticompetitive conduct lacking a plausible efficiency jus-tification. The term “naked” is often applied to collusive agreements amongrivals to fix prices, divide markets, or otherwise harm competition that cannotplausibly be justified as efficient.79 The term “plain exclusion” will be used todescribe the comparable exclusionary conduct: anticompetitive exclusionlacking a plausible efficiency justification.80 “Cheap exclusion” is a type ofplain exclusion, namely plain exclusion that is also inexpensive for the ex-cluding firms to implement.81

It is commonplace today that agreements among rivals (which more com-monly threaten collusive rather than exclusionary harms), when reviewedunder Sherman Act Section 1, are analyzed under the rule of reason through

ers, or as monopolization or attempts to monopolize, those categories can also be employed toattack collusive conduct, and the legal categories in which collusive cases tend to be framed,including horizontal agreements, can also be employed to attack exclusionary conduct. Within adoctrinal category, moreover, the legal rule generally does not differ depending on whether thealleged conduct is collusive or exclusionary. The rules governing group boycotts may be anexception, however. The Supreme Court’s collusive group boycott decision in FTC v. SuperiorCourt Trial Lawyers Association (SCTLA) treated that conduct as tantamount to price fixingamong rivals. 493 U.S. 411, 423 (1990). The SCTLA majority did not make reference to theCourt’s then-recent exclusionary group boycott decision, Northwest Wholesale Stationers, Inc. v.Pacific Stationery and Printing Co., 472 U.S. 284 (1985). Stationers appears to demand a moreextensive showing (perhaps including proof of market power) before applying a per se rule toinvalidate the conduct than is required for horizontal price fixing—though it is not possible tosay more than “appears” and “perhaps” because Stationers does not clearly delineate the ele-ments of the per se rule it applies. In the wake of Stationers, the lower courts have grappledinconclusively with the issue. See 1 ANTITRUST LAW DEVELOPMENTS, supra note 36, at 492–93.

79 See, e.g., White Motor Co. v. United States, 372 U.S. 253, 263 (1963) (“Horizontal territo-rial limitations . . . are naked restraints of trade with no purpose except stifling of competition.”).

80 The seemingly analogous term “naked exclusion” was not adopted because that phrase isused in the economics literature to describe a particular economic model. See Eric B. Rasmusen,J. Mark Ramseyer & John S. Wiley, Jr., Naked Exclusion, 81 AM. ECON. REV. 1137 (1991). Butsee Jonathan M. Jacobson, Exclusive Dealing, “Foreclosure,” and Consumer Harm, 70 ANTI-

TRUST L.J. 311, 360 (2002) (referring to “naked” exclusion); Krattenmaker & Salop, supra note9, at 227 (same). Plain exclusion in a Sherman Act Section 2 setting has been referred to as “noefficiency justification” monopolization. Harry First, The Case for Antitrust Civil Penalties, 76ANTITRUST L.J. 127, 160 (2009).

81 Susan A. Creighton, D. Bruce Hoffman, Thomas G. Krattenmaker & Ernest A. Nagata,Cheap Exclusion, 72 ANTITRUST L.J. 975, 977 (2005); Patricia Schultheiss & William E. Cohen,Cheap Exclusion: Role and Limits (Jan. 14, 2009) (unpublished manuscript), available at http://www.ftc.gov/os/sectiontwohearings/docs/section2cheapexclusion.pdf. The concept of cheap ex-clusion was developed in part as a guide to the enforcement agencies in allocating investigativeresources, and incorporated the expense of implementation on the view that it would be related tothe likelihood of uncovering anticompetitive exclusion.


an analysis that can be structured or truncated using quick look or burden-shifting approaches.82 In consequence, a horizontal restraint can be con-demned without a comprehensive analysis of its effects on competition ifthree elements are demonstrated: (a) an agreement among rivals,83 (b) certainfacts suggesting the likelihood of harm to competition; and (c) the absence ofa plausible efficiency justification for the agreement at issue. The second ele-ment may be satisfied by showing that the conduct falls in a traditional per secategory (price fixing or market division),84 by showing that anticompetitiveeffect is intuitively obvious based on facial analysis of the agreement,85 or(with retrospective conduct) through actual effects evidence demonstratingthat competition has been harmed.86

82 See Andrew I. Gavil, Moving Beyond Caricature and Characterization: The Modern Ruleof Reason in Practice, 85 S. CAL. L. REV. 733, 744–53 (2012); see also ANTITRUST LAW IN

PERSPECTIVE, supra note 25, at 206 (plaintiffs seek courts to truncate the rule of reason review ofhorizontal restraints in order “to condemn conduct without detailed analysis of market power andlikely effects” when the conduct “is facially objectionable or has actual adverse effects,” and todo so when the conduct “would be in a traditional per se category but for plausible efficiencies,and on review the efficiencies do not actually appear substantial”). Courts may also considertruncating the rule of reason review of horizontal restraints in order to exculpate conduct whendefendants collectively have a low market share. Id. In the context of burden shifting, however,the latter possibility would presumably be considered only after both plaintiff and defendant havesatisfied their initial burdens of production.

83 Application of Sherman Act Section 1 is predicated on proof of agreement; the rules dis-cussed in this paragraph govern the analysis of agreements among (horizontal) competitors. SeeSherman Act, 15 U.S.C. § 1. The agreement element often goes undisputed, but if an agreementamong rivals must be inferred from circumstantial evidence, that element may be difficult toassess. See generally Baker, supra note 25; Louis Kaplow, Direct vs. Communications-BasedProhibitions on Price-Fixing, 3 J. LEGAL ANALYSIS 449 (2011). Antitrust law could in theorycondemn collusive conduct on a truncated basis without proof of agreement among rivals if theother elements of the truncated rule are present, but the limited experience with identifying collu-sive effects from unilateral conduct after Ethyl and from vertical agreements after GTE Sylvaniaand Leegin offers little guidance as to whether and when courts would do so. See generallyLeegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877 (2007) (resale price mainte-nance); Cont’l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) (non-price restraints); E.I. duPont de Nemours & Co. v. FTC, 729 F.2d 128 (2d Cir. 1984) (Ethyl).

84 See, e.g., Palmer v. BRG of Ga., Inc., 498 U.S. 46 (1990) (horizontal market division);United States v. Trenton Potteries Co., 273 U.S. 392 (1927) (horizontal price fixing). The Su-preme Court has also recognized that a collusive group boycott is tantamount to horizontal pricefixing. FTC v. Superior Ct. Trial Lawyers Ass’n, 493 U.S. 411, 436 n.19 (1990).

85 See, e.g., PolyGram Holding, Inc. v. FTC, 416 F.3d 29, 36 (D.C. Cir. 2005) (“If, based uponeconomic learning and the experience of the market, it is obvious that a restraint of trade likelyimpairs competition, then the restraint is presumed unlawful . . . .”). The court went on to findthis criterion satisfied by an agreement between joint venturers to restrain price cutting and ad-vertising with respect to products not part of the joint venture. Id.

86 E.g., FTC v. Ind. Fed’n of Dentists, 476 U.S. 447 (1986); NCAA v. Bd. of Regents of theUniv. of Okla., 468 U.S 85 (1984). But see Republic Tobacco Co. v. N. Atl. Trading Co., 381 F.3d 717, 737 (7th Cir. 2004) (noting that even though Indiana Federation of Dentists does notallow a plaintiff to dispense entirely with market definition by proffering actual effects evidence,a plaintiff must still show the “rough contours” of a market and that defendant commands asubstantial share). In a retrospective exclusion case, for example, proof that prices rose after arival was excluded might count as actual effects evidence. (The probative value of actual effects


This approach is typically implemented today through a burden-shiftingframework that has been developed by the lower courts in agreement casesalleging collusive effects.87 A plaintiff must satisfy an initial burden of pro-duction by demonstrating that competition has been or likely will be harmed.88

If the plaintiff makes a satisfactory initial showing, the burden of productionshifts to defendants to identify a plausible business justification. If defendantdoes so, the plaintiff, on whom the burden of persuasion rests, must proveunreasonableness by showing that the harm to competition is not dissipated oreliminated by the benefit to competition,89 or that defendant had a practicalless-restrictive alternative for achieving the benefits with less harm tocompetition.90

The burden-shifting framework implies that the rule of reason review ofallegedly collusive horizontal agreements can be truncated relative to the waya court would proceed under the comprehensive rule of reason in two senses.91

evidence can be contested—in this example, perhaps, with evidence that non-excluded firmsexperienced an independent increase in marginal cost of sufficient magnitude to explain the priceincrease.)

87 See, e.g., PolyGram Holding, 416 F.3d at 36 (D.C. Cir. 2005); Law v. NCAA, 134 F.3d1010, 1019 (10th Cir. 1998); see also Andrew I. Gavil, Burden of Proof in U.S. Antitrust Law, in1 ISSUES IN COMPETITION LAW AND POLICY, supra note 52, at 125, 145–48; Gavil, supra note 82,at 758–60 (discussing burden shifting and the rule of reason).

88 The plaintiff may meet this burden with any of the limited factual showings of harm tocompetition that would provide a basis for truncated or quick look condemnation: that the agree-ment falls in a traditional per se category, that harm is intuitively obvious, or that harm hasalready occurred (actual effects evidence). Because the plaintiff also has the option of provingunreasonableness through a comprehensive rule of reason review, the plaintiff can also satisfy itsinitial burden with a more detailed demonstration of harm to competition based on an analysis ofa wider range factors such as defendant market power or the actual effects of the agreement asimplemented—in which case the plaintiff’s initial burden of production would merge with itsultimate burden of persuasion. In practical application under Sherman Act Section 1, most plain-tiffs have failed to satisfy their initial burden. See Michael A. Carrier, The Real Rule of Reason:Bridging the Disconnect, 1999 BYU L. REV. 1265, 1293 [hereinafter Bridging the Disconnect];Michael A. Carrier, The Rule of Reason: An Empirical Update for the 21st Century, 16 GEO.MASON L. REV. 827, 831–32 (2009) [hereinafter Empirical Update] .

89 Courts routinely describe the unstructured reasonableness inquiry in terms of balancing ben-efits and harms, but in practice they almost never actually balance. See Carrier, Bridging theDisconnect, supra note 88, at 1293; Carrier, Empirical Update, supra note 88, at 831–32. Ac-cordingly, following a suggestion of Prof. Andrew Gavil, in this article I describe reasonablenessreview as evaluating whether the benefits “dissipate or eliminate” the harms rather than as “bal-ancing” or “weighing” harms against benefits. If a court were to permit efficiency benefits in onemarket to justify conduct that harmed competition in a different market, however, it would bedifficult to interpret that process other than as balancing.

90 For one court’s statement of this framework, see Law v. NCAA, 134 F.3d at 1019; and seealso Gregory v. Fort Bridger Rendezvous Association, 448 F.3d 1195, 1205 (10th Cir. 2006)(reaffirming Law framework). For a collection of cases, see 1 ANTITRUST LAW DEVELOPMENTS,supra note 36, at 62 n.353.

91 This article uses the phrases “comprehensive rule of reason,” “unstructured rule of reason,”and “full blown rule of reason” interchangeably to refer to the type of wide-ranging analysisundertaken in Board of Trade. See generally Bd. of Trade of City of Chicago v. United States,246 U.S. 231 (1918).


First, a plaintiff may satisfy its initial burden without undertaking a detailedmarket analysis (which would require defining markets, analyzing marketshares, evaluating entry conditions, and the like) by relying instead upon cate-gorization of the agreement, facial analysis of the agreement, or actual effectsevidence. In addition, harm to competition may be inferred from the limitedshowing required to satisfy the plaintiff’s initial burden combined with theabsence of plausible efficiencies, without need for further analysis.

The courts appear to be developing a structured approach for evaluatinganticompetitive exclusion that is similar to the approach they apply to evalu-ate anticompetitive collusion. As with alleged collusive harms, allegations ofanticompetitive exclusion are generally tested under the rule of reason acrossdoctrinal categories.92 Exclusive dealing allegations are evaluated for theirreasonableness, whether challenged under the Sherman Act or the ClaytonAct.93 Vertical agreements, which could result in exclusion, are reviewedunder the rule of reason regardless of whether they involve price or non-priceterms.94 Tying and exclusionary group boycotts are evaluated under the rule ofreason if a per se rule does not apply.95 The exclusionary conduct element ofthe monopolization offense is reviewed in a burden-shifting framework simi-lar to the approach now applied to evaluate the reasonableness of conductunder Sherman Act Section 1.96

92 Many economic factors relevant to showing that exclusion and collusion have harmed com-petition, discussed below in Parts III and IV, would be relevant when applying the unstructuredrule of reason, but they would not all be relevant if the reasonableness review is truncated.

93 Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 44–45 (1984) (O’Connor, J, concur-ring) (Sherman Act); Omega Envtl., Inc. v. Gilbarco, Inc., 127 F.3d 1157, 1162 (9th Cir. 1997)(Clayton Act); U.S. Healthcare, Inc. v. Healthsource, Inc., 986 F.2d 589, 593 (1st Cir. 1993)(Sherman Act); see Jacobson, supra note 80, at 363 (noting that exclusive dealing analysis hasbeen “freed . . . to conform to more general analysis of trade restraints under the rule of reason”).

94 Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877 (2007) (resale price main-tenance); Cont’l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) (non-price restraints).

95 Jefferson Parish, 466 U.S. at 29 (tying); United States v. Microsoft Corp., 253 F.3d 34, 84(D.C. Cir. 2001) (tying); see Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co.,472 U.S. 284, 297 (1985) (exclusionary group boycott); United States v. Visa U.S.A., Inc., 344F.3d 229 (2d Cir. 2003) (analyzing conduct tantamount to an exclusionary group boycott underthe rule of reason).

96 Compare Microsoft, 253 F.3d at 58–59, with PolyGram Holding, Inc. v. FTC, 416 F.3d 29(D.C. Cir. 2005), and Law v. NCAA, 134 F.3d 1010 (10th Cir. 1998). Cf. United States v.Standard Oil Co., 221 U.S. 1, 61–62 (1911) (noting that the rule of reason applies to the analysisof conduct under both Sherman Act Section 1 and Sherman Act Section 2). The reasonablenessanalysis of monopolization is structured further when price-cutting is the alleged exclusionaryact, as predatory pricing requires proof of below-cost pricing and an assessment of the price-cutter’s prospects for recouping the costs of below-cost pricing through the later exercise ofmonopoly power. See Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209(1993) (applying Sherman Act principles to Robinson-Patman Act decision). The recoupmentinquiry can be understood as assessing the profitability of the alleged anticompetitive strategy,and thus evaluating whether the excluding firms can solve the third “exclusion problem” dis-cussed below in Part IV.


The courts have also arguably begun to develop an approach for truncatingthe rule of reason review of exclusionary conduct across legal categories,much as they have come to do with collusive horizontal agreements. Synthe-sizing the leading cases, exclusionary conduct may be found unreasonabletoday without a comprehensive analysis of the nature, history, purpose, andactual or probable effect of the practice in the presence of two additional ele-ments: (i) if the excluding firms have foreclosed competition from all actualor potential rivals other than insignificant competitors,97 and (ii) if the exclu-sionary conduct lacks a plausible efficiency justification.98 The need to iden-tify the excluding firms’ rivals may call for at least an informal marketdefinition,99 but this predicate would not undermine the benefits of truncationin reducing transaction costs and providing guidance when market definitionis not difficult,100 and may not fully undermine those benefits even if market

97 Aspen Skiing v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985) (discussing a domi-nant firm that excluded its only competitor); Microsoft, 253 F.3d at 55 (noting that exclusionaryconduct protected the “applications barrier to entry” that insulated the dominant firm from com-petition from current and potential rivals); United States v. Dentsply Int’l., Inc., 399 F.3d 181 (3dCir. 2005) (discussing a dominant firm that foreclosed its rivals from access to dealers, andnoting that while this exclusionary method did not cover two small rivals that sold directly to theultimate customers, their alternative method of distribution was less effective). In Lorain JournalCo. v. United States, 342 U.S. 143 (1951), the dominant daily newspaper excluded a rival radiostation from the advertising market but did not exclude another competitor, a weekly newspaper.The Supreme Court and the district court appear to have treated the small weekly newspaper asan insignificant market participant, in which case the exclusionary conduct foreclosed only thesole significant rival, but these decisions could instead be read to have defined the market nar-rowly to exclude that firm as a participant. United States v. Lorain Journal Co., 92 F. Supp. 794,796–97 (N.D. Ohio 1950), aff’d, 342 U.S. 143 (1951).

98 See United States v. Visa U.S.A., Inc., 344 F.3d 229 (2d Cir. 2003) (government prevailedby showing harm to competition and the absence of procompetitive benefits, though the inquiryinto competitive harm was wide-ranging); Gavil, supra note 14, at 27–28 (noting that plaintiffsare most likely to succeed in proving exclusionary violations under Sherman Act Section 2 whenthe harm to competition or defendant’s market power are obvious and the defendant lacks aplausible business justification); see also Mark S. Popofsky, Defining Exclusionary Conduct:Section 2, the Rule of Reason, and the Unifying Principle Underlying Antitrust Rules, 73 ANTI-

TRUST L.J. 435, 445 (2006) (the Microsoft framework for evaluating exclusionary conduct underSherman Act Section 2 “is virtually indistinguishable from the test courts employ under Section1’s rule of reason”).

99 See Republic Tobacco Co. v. N. Atlantic Trading Co., 381 F. 3d 717, 737 (7th Cir. 2004)(noting that the approximate magnitude of market shares may be assessed after proving the“rough contours” of a market). By contrast, the second element in the truncated reasonablenessreview of collusive conduct does not require market definition. If truncated condemnation ofexclusionary conduct is instead based on actual evidence of anticompetitive effect—an openpossibility, see infra notes 129–31 and accompanying text—market definition would presumablynot be required.

100 Market definition appears to have been uncontroversial in a number of pro-plaintiff exclu-sion decisions. See, e.g., Lorain Journal, 342 U.S. 143 (mass dissemination of news and adver-tising, both of a local and national character, in Lorain, Ohio); Dentsply, 399 F.3d 181 (sale ofprefabricated artificial teeth in the United States); Conwood Co. v. U.S. Tobacco Co., 290 F.3d768 (6th Cir. 2002) (moist snuff in the United States).


definition is strongly contested.101 Moreover, it may not be necessary to iden-tify with specificity every significant foreclosed rival to determine that allsuch rivals were excluded.102

The D.C. Circuit’s analysis of monopolization in its en banc Microsoft de-cision provides the most detailed articulation of this approach.103 The monopo-lization offense has two doctrinal elements: monopoly power andexclusionary conduct (acquiring or maintaining that power through anticom-petitive means).104 To identify the second element, exclusionary conduct, thecircuit court adopted a burden-shifting framework consistent with “a centuryof case law on monopolization[.]”105 Under its scheme, plaintiff initially estab-lishes a prima facie case by demonstrating anticompetitive effect.106 If plaintiffsuccessfully does so, the defendant may proffer a non-pretextual procompeti-tive justification for its conduct, which shifts the burden back to the plaintiffto rebut that claim.107 If the justification stands unrebutted, the plaintiff mustdemonstrate that the anticompetitive harm of the conduct outweighs theprocompetitive benefit.108 In applying this framework, proof that the monopo-list has foreclosed all actual or potential rivals would undoubtedly be suffi-cient to establish plaintiff’s prima facie case,109 after which defendant’sinability to identify a plausible efficiency justification would suffice to provethat the conduct is exclusionary and thus to condemn it as a violation of Sher-man Act Section 2.

Truncated condemnation on a similar basis appears possible across most ifnot all of the disparate legal categories in which exclusionary conduct allega-tions may be evaluated, including attempt to monopolize,110 monopoliza-

101 Problems that arise in defining markets in exclusion settings are discussed in Jonathan B.Baker, Market Definition: An Analytical Overview, 74 ANTITRUST L.J. 129, 166–73 (2007).

102 Applications of Comcast Corporation, General Electric Company, and NBC Universal, Inc.,for Consent to Assign Licenses and Transfer Control of Licensees, Memorandum Opinion andOrder, 26 FCC Rcd. 48 (2011), available at transition.fcc.gov/FCC-11-4.pdf; see also UnitedStates v. Microsoft Corp., 253 F.3d 34, 79–80 (D.C. Cir. 2001) (inferring causal link betweendefendant’s anticompetitive conduct and maintenance of its monopoly from the exclusion ofnascent competitive threats, but not identifying each excluded potential rival).

103 See Microsoft, 253 F.3d at 78–80.104 United States v. Grinnell Corp., 384 U.S. 563, 570–71 (1966). Microsoft describes the sec-

ond element as “exclusionary conduct.” Microsoft, 253 F.3d at 58.105 Microsoft, 253 F.3d at 58.106 Id. at 58–59.107 Id. at 59.108 Id.109 Market definition may be the predicate for identifying rivals. See supra notes 99–102 and

accompanying text. In Microsoft, the court identified a market in the context of evaluating themonopoly power element of the monopolization offense. See Microsoft, 253 F.3d at 50–56.

110 See Lorain Journal Co. v. United States, 342 U.S. 143 (1951) (dominant daily newspaperexcluded rival radio station from advertising market). Neither the Supreme Court nor the districtcourt made clear whether advertising competition from a small weekly newspaper in the same


tion,111 exclusionary group boycott,112 non-price vertical restraints,113 andexclusive dealing.114 The case law establishing the truncated approach to rea-sonableness review of exclusionary conduct does so more clearly for monopo-lization, the legal category addressed by Microsoft’s burden-shiftingframework, than for violations that would fall in other legal categories. Forthis reason, the cases could be read as establishing a burden-shifting approachto truncation only for the review of monopolization allegations. That narrowreading would be inconsistent with the broad trend in antitrust elevating con-cepts over categories discussed above in Part I.115 Hence, modern courtswould likely look across doctrinal pigeonholes for guidance in evaluating ex-clusionary conduct claims, and adopt a common framework for doing so.

The specifics of the structured approach are less evident in the exclusioncontext than the collusive one because many more collusion cases have beencondemned after truncated review than exclusion cases.116 It is nevertheless

city was ruled out by defining the market to exclude that firm as a participant—in which case theexclusionary conduct at issue foreclosed competition from all rivals—or whether it treated thatfirm as an insignificant market participant, in which case the exclusionary conduct foreclosedonly the sole significant rival. United States v. Lorain Journal Co., 92 F. Supp. 794, 796–97 (N.D. Ohio 1950), aff’d, 342 U.S. 143 (1951).

111 Aspen Skiing v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985); Microsoft, 253 F.3dat 64–78. Cf. Jonathan B. Baker, Promoting Innovation Competition Though the Aspen/KodakRule, 7 GEO. MASON L. REV. 495, 496 (1999) (explaining that the Supreme Court had estab-lished a truncated legal rule finding a violation of Sherman Act Section 2 when a monopolistexcludes rivals by restricting a complementary or collaborative relationship without an adequatebusiness justification).

112 See Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284, 298(1985) (remanding for review under a legal rule creating the possibility of truncated condemna-tion). The Stationers Court used the term “per se rule,” but, in contrast to the traditional per serules employed in the analysis of horizontal restraints, conditioned application of its rule on up tothree elements including defendant market power. See supra note 78. For this reason, its ap-proach is better thought of as describing a truncated or structured inquiry. Accord, Toys “R” Us,Inc. v. FTC, 221 F.3d 928, 936 (7th Cir. 2000).

113 See Graphic Prods. Distribs., Inc. v. ITEK Corp., 717 F.2d 1560, 1583 (11th Cir. 1983)(affirming jury verdict for plaintiff upon proof of defendant market power absent evidence thatthe restraints were reasonably necessary to achieve a legitimate business purpose); Eiberger v.Sony Corp., 622 F.2d 1068 (2d Cir. 1980) (affirming unjustified exclusion condemned withoutfurther inquiry into defendant market power); Gavil, supra note 14, at 8 n.29 (“[Continental T.V.,Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977),] strongly implies that if a plaintiff can demon-strate that its supplier possesses market power, the burden of production should shift to thedefendant to justify its conduct.”). Although non-price vertical restraints can be subject to trun-cated condemnation, rule of reason litigation of such agreements almost always ends with defen-dant prevailing. See HOVENKAMP, supra note 36, § 11.6b; ANTITRUST LAW IN PERSPECTIVE,supra note 25, at 369 (Sidebar 4-1: Dealer Relations After Sylvania).

114 Cf. United States v. Dentsply Int’l, Inc., 399 F.3d 181 (3d Cir. 2005) (analyzing exclusivedealing conduct as monopolization).

115 See supra notes 25–27 and accompanying text.116 The small number of recent exclusion decisions consistent with the synthesized rule does

not mean that there is no such rule; it more likely shows that there are fewer litigated exclusion-ary conduct claims than collusive conduct claims in the first place, or that the conduct in suchcases is less frequently justified with a plausible efficiency claim. The enforcement agencies


evident from the synthesized rule that as the legal rules governing exclusionand collusion evolve, they are converging on an approach that is more likelyto find unlawful conduct lacking a plausible efficiency justification, regardlessof whether the anticompetitive effects are exclusionary or collusive—thusdemonstrating that antitrust’s doctrinal rules evaluate collusion and exclusionin a similar way, and do not tilt the scales to downplay exclusion.

C. REFINING THE TRUNCATED RULE GOVERNING EXCLUSION

The courts have not explicated as doctrine the synthesized rule for trun-cating the reasonableness review of exclusionary conduct set forth above.117 Inconsequence, many questions about truncated condemnation of exclusionaryconduct remain open for future refinement, including the following six.118

First, if some rivals are not excluded, what showing is required to demon-strate their competitive insignificance? As an economic matter, a firm or firmswould be insignificant for this purpose if, given the cost and difficulty ofoutput expansion, it or they (collectively) would not increase sales sufficientlyto undermine the post-exclusion exercise of market power. Courts that havetreated non-excluded firms as insignificant have not experienced difficultyconcluding that they have high costs of output expansion or low marketshares,119 but these factual determinations will not always be easy to make.120

more frequently challenge naked collusion than plain exclusion. See infra note 237 and accompa-nying text (discussing relative frequency of exclusion and collusion cases).

117 An earlier work by the present author described a truncated legal rule established by theSupreme Court in two monopolization decisions: Aspen and Kodak. Aspen Skiing v. AspenHighlands Skiing Corp., 472 U.S. 585 (1985); Eastman Kodak Co. v. Image Technical Servs.,Inc., 504 U.S. 451 (1992); Baker, supra note 111, at 496. Under that rule, Sherman Act Section 2is violated when a monopolist excludes rivals by restricting a complementary or collaborativerelationship without an adequate business justification. This discussion updates that prior analy-sis to reflect more recent precedent. Most importantly, a number of questions about the elementsof the rule that seemed open in 1999, see Baker, supra, at 503–05, have been addressed throughthe analytical framework set forth in Microsoft, United States v. Microsoft Corp., 253 F.3d 34(D.C. Cir. 2001) (en banc). The generality of the Microsoft framework also suggests, contrary tothe cautious interpretation of the cases offered in 1999, that truncation does not turn on themeans of exclusion (that is, it would not be limited to exclusionary conduct that takes the form ofa restriction to a prior complementary or collaborative relationship). See id.

118 In addition, it is an open question whether truncated condemnation under the rule of reasoncan be applied to exclusionary conduct undertaken by or associated with the creation of jointresearch or production ventures, including voluntary standards development organizations, asthese types of joint ventures “shall be judged on the basis of its reasonableness, taking intoaccount all relevant factors affecting competition . . . .” 15 U.S.C. § 4302.

119 A low market share may indicate competitive insignificance in this context because smallfirms often have difficulty expanding output inexpensively, as would be the case if they are smallmainly for reasons other than aggressive competition by the excluding firms. Even if productioncosts are low and do not increase with output, for example, the costs of identifying and market-ing to new customers often increase as output rises.

120 But see supra note 100 (providing examples from exclusionary conduct cases in whichcourts did not find market definition difficult).


If competitive insignificance is not easily determined, the truncated rule couldlose its administrability advantage over unstructured rule of reason review,leading a court to prefer not to decide the matter before it on a quick look.

Second, if the excluding firm or firms have market power (most likelydemonstrated through market shares) can a court infer that all significant ri-vals, actual or potential, have been excluded or likely would be excluded fromevidence that one such rival has been excluded? That is, if the excluding firmshave market power and are able to foreclose one rival, is it reasonable topresume they have the ability and incentive to foreclose all rivals? The Fed-eral Trade Commission answered this question in the affirmative in a decisioncondemning an exclusionary group boycott.121 If its answer is generally ac-cepted, truncated condemnation could be undertaken without identifyingevery significant rival and proving that each has been excluded or likelywould be excluded.122 Instead, exclusionary conduct would be condemnedwithout full rule of reason review on a showing that one or more rivals wereexcluded, the excluding firms possess market power, and the exclusionaryconduct at issue had no plausible efficiency justification.

Third, in a prospective exclusion case, if the exclusionary conduct fore-closes all actual and potential rivals, and has no business justification, can itbe condemned without proof that the excluding firms previously had marketpower—particularly regardless of the excluding firm’s market share?123 Theobvious logic of the inference created by the synthesized rule set forth

121 Toys “R” Us, Inc., 126 F.T.C. 415, 590–608 (1998), aff’d, 221 F.3d 928 (7th Cir. 2000); seeMultistate Legal Studies, Inc. v. Harcourt Brace Jovanovich Legal and Prof’l Publ’ns, Inc., 63F.3d 1540, 1549, 1555–56 (10th Cir. 1995) (explaining that evidence of defendant market powerand entry barriers suggests a triable issue of fact concerning recoupment in a predatory pricingcase). Moreover, if harm to competition can be inferred from proof of market power and theabsence of efficiencies, it is an open question what market share would be sufficient to satisfy therule. In Toys, the FTC applied the rule to a firm it found to have a market share of more than 30percent in the areas in which it did business and between 40 percent and 50 percent in manycities. Toys, 126 F.T.C. at 597–99. The risk of a false positive—here the risk of wrongly infer-ring that non-excluded rivals would be unable to counteract the harm to competition by collec-tively expanding output—may be greater the lower the market share threshold.

122 See Christine A. Varney, A Post-Leegin Approach to Resale Price Maintenance Using aStructured Rule of Reason, ANTITRUST, Fall 2009, at 22, 24–25 (proposing that a plaintiff pro-ceeding on a manufacturer or retailer exclusion theory of harm from resale price maintenancecan satisfy its prima facie case, thus shifting a burden to defendant to prove efficiencies, byshowing: that defendant has market power, that the resale price maintenance arrangements resultin substantial foreclosure, and that exclusionary effects are plausible); cf. Nw. Wholesale Station-ers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284 (1985) (truncated condemnation rulelooks only to exclusion, market power, and the absence of efficiencies).

123 Market power in exclusion cases can also be demonstrated by actual effects evidence. E.g.,Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 477 (1992) (noting thatmarket power may be inferred from direct evidence that prices rose and rivals were excluded);United States v. Microsoft Corp., 253 F.3d 34, 51, 57 (D.C. Cir. 2001) (holding on alternativegrounds of monopoly power); ReMax Int’l, Inc. v. Realty One, Inc., 173 F.3d 995 (6th Cir.1999); see also Geneva Pharms. Tech. Corp. v. Barr Labs. Inc., 386 F.3d 485, 500 (2d Cir. 2004)


above—a firm clearly can exercise market power by excluding all its rivals nomatter how small its prior share—implies that proof of pre-existing marketpower should be unnecessary for truncated condemnation.124 Moreover, thatoutcome is consistent with the rule governing attempted monopolization,which requires proof of a “dangerous probability” of achieving monopolypower rather than pre-existing monopoly power,125 and the case law establish-ing that a monopoly obtained through the fraudulent acquisition of a patentviolates Sherman Act Section 2.126 Given the historical importance of defen-dant market share in evaluating allegations of anticompetitive exclusion, evenoutside of Sherman Act Section 2,127 however, it is possible that a court couldnevertheless require proof of excluding firm market power before truncatingits reasonableness review.128

Fourth, is truncated condemnation available in a retrospective exclusioncase without proof of market power but with a showing of actual anticompeti-tive effects? This approach would follow the logic of a quick look methodol-ogy established in collusive agreement cases,129 and was endorsed for

(showing of adverse effects insufficient on the facts of the case); Tops Mkts., Inc. v. QualityMkts., Inc., 142 F.3d 90, 97 (2d Cir. 1998) (same).

124 “A firm that seeks to create a monopoly by dynamiting its competitor’s plants does not needmarket power—only a saboteur and a match.” See HOVENKAMP, supra note 36, § 6.5.

125 Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 455 (1993). Market shares too low toprove market power may be sufficient to show “dangerous probability.” 1 ANTITRUST LAW DE-

VELOPMENTS, supra note 36, at 323.126 Walker Process Equip. Inc. v. Food Mach. & Chem. Corp., 382 U.S. 172, 173–74 (1965).

To similar effect, suppose a firm manipulates a standard-setting process through deception toensure that the standard incorporates its intellectual property, giving it the potential to exercisemarket power by asserting intellectual property rights. This showing combined with proof thatthe exclusionary conduct has no legitimate business justification would likely be sufficient toprove harm to competition if there is no practical way to compete without complying with thestandard. See Rambus Inc. v. FTC, 522 F.2d 456, 467–68 (D.C. Cir. 2008) (finding no antitrustviolation because the proof of exclusion was insufficient on the facts of the case); Note, Decep-tion as an Antitrust Violation, 125 HARV. L. REV. 1235, 1251–54 (2011) (advocating rule per-mitting court to find monopolization when a monopolist’s deceptive act was reasonably capableof contributing to monopoly power, and to find an agreement to deceive unreasonable if it cre-ates a significant anticompetitive effect).

127 See, e.g., 1 ANTITRUST LAW DEVELOPMENTS, supra note 36, at 216 (“Since the early 1970s,judicial decisions [in exclusive dealing cases] have established a virtual safe harbor for marketforeclosure of 20 percent or less.”).

128 See Jacobson, supra note 80, at 365 (noting that in the reasonableness analysis of exclusivedealing within a burden-shifting framework allowing for truncated condemnation, a plaintiffmust prove defendant market power to satisfy its initial burden).

129 FTC v. Ind. Fed’n of Dentists, 476 U.S. 447 (1986); NCAA v. Bd. of Regents of the Univ.of Okla., 468 U.S 85 (1984).


anticompetitive exclusion by the Federal Trade Commission,130 but has beenquestioned by the Seventh Circuit.131

Fifth, how will the synthesized rule for truncating the reasonableness re-view of exclusionary conduct be harmonized with the rules establishing be-low-cost pricing and recoupment as elements of the predatory pricingoffense?132 Truncated condemnation is unlikely to be available in such casesbecause one or more of the many competitive justifications for low priceswould typically appear plausible,133 rather than because these elements arepart of the offense. (On similar reasoning. truncated condemnation is also un-likely to be available when the exclusionary conduct involves the introductionof a (non-sham) product design improvement.)134

Sixth, what business justifications for exclusionary conduct are cogniza-ble?135 For example, can defendants justify exclusionary conduct on theground that the opportunity to charge monopoly prices induces the excludingfirms to invest in developing or marketing innovative products or productionprocesses?136 This argument would seem to be “a defense based on the as-

130 See, e.g., Toys “R” Us, Inc., 126 F.T.C. 415, 608 (1998), aff’d, 221 F.3d 928 (7th Cir.2000).

131 In Republic Tobacco Co. v. North Atlantic Trading Co., 381 F. 3d 717, 737–38 (7th Cir.2004), the Seventh Circuit declined to allow the plaintiff to dispense with market definition byproffering actual effects evidence in a vertical exclusion case. Although this position had seem-ingly been rejected by the Supreme Court when previously adopted by the same circuit court, seeIndiana Federation of Dentists, 476 U.S. at 465–66, rev’g 745 F.2d 1124 (7th Cir. 1984), theappellate court in Republic Tobacco interpreted Indiana Federation of Dentists to require plain-tiff to show the “rough contours” of a market and that the defendant commands a substantialshare, Republic Tobacco, 381 F.3d at 737. That Seventh Circuit took a different view of IndianaFederation of Dentists in an earlier exclusionary group boycott case not discussed in RepublicTobacco. See Wilk v. AMA, 895 F.2d 352, 360 (7th Cir. 1990) (alternative holding).

132 Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209 (1993).133 See Baker, supra note 46, at 587–89 (surveying procompetitive explanations for prices that

might appear to below costs); see also John E. Lopatka & William H. Page, ‘Obvious’ ConsumerHarm in Antitrust Policy: The Chicago School, the Post-Chicago School and the Courts, inPOST-CHICAGO DEVELOPMENTS IN ANTITRUST LAW, supra note 6, at 129, 132–36 (noting thatcourts tend to have sympathy for conduct conferring short-run consumer benefits and hostilityfor conduct conferring short-run consumer harm, even when careful economic analysis mightsupport a decision otherwise).

134 Allied Orthopedic Appliances Inc. v. Tyco Health Care Grp., 592 F.3d 991, 1000 (9th Cir2010). See generally Alan Devlin & Michael Jacobs, Anticompetitive Innovation and the Qualityof Invention, 27 BERKELEY TECH. L.J. 1 (2012) (discussing the relationship between acts ofinvention and antitrust violations).

135 In addition to the cognizability issue highlighted in the text, other issues familiar from otherantitrust contexts may include the treatment of cost savings that benefit sellers but are not passedthrough to buyers, and whether or when to count efficiencies that benefit buyers in markets otherthan the market in which the harm to competition occurs.

136 To similar effect, one commentator raises the possibility that a monopolist’s exclusion ofrivals from access to one side of its two-sided platform could be justified by its success inmaking the platform more effective at attracting buyers on the other side. See Jacobson, supranote 80, at 361 (“[O]ne can imagine a nonfrivolous (albeit weak) argument on behalf of the


sumption that competition itself is unreasonable,” and thus ruled out by theholding of National Society of Professional Engineers,137 but dicta in Trinkomight appear to call that conclusion into question.138

In a general sense, the truncated rule of law explicated above allows con-demnation of exclusion as anticompetitive without comprehensive reasonable-ness upon a showing of three elements: (a) exclusionary conduct, (b) factssuggesting the likelihood of harm to competition; and (c) the absence of aplausible efficiency justification for the exclusionary conduct. The second ele-ment would be satisfied by the exclusion of all actual and potential rivals(other than insignificant ones); the open questions raise the possibility that itmay also be satisfied in other ways, particularly by excluding-firm marketpower combined with the ability to foreclose at least one rival, or by actualanticompetitive effects.

Describing the truncated rule for exclusion this way emphasizes its struc-tural similarity to the truncated rule for horizontal restraints applied to collu-sion, which, as previously discussed,139 establishes a violation on a showing ofthree similar general elements: (a) an agreement among rivals, (b) facts sug-gesting the likelihood of harm to competition; and (c) the absence of a plausi-ble efficiency justification for the agreement. If the predicates for quick lookcondemnation are not satisfied, the conduct is subject to unstructured rule ofreason review in each case. As a formal matter, therefore, antitrust’s doctrinalrules treat plain exclusion and naked collusion comparably; they do not confera more relaxed scrutiny on exclusionary conduct.

The structured doctrinal rules for exclusion and collusion both require theabsence of a plausible efficiency justification as a predicate for truncated con-demnation, but the rules differ in their first two elements. Their initial ele-ments are obviously not the same: the exclusion rule is predicated onexclusionary conduct while the collusion rule is predicated on collusive con-duct.140 As will be examined in detail in Part IV, moreover, the rules alsodiffer in their second element. The facts suggesting that the collusive conductis anticompetitive, such as price fixing or market division, differ from the

Lorain Journal that the value of the newspaper as an advertising medium might be diluted if thesame messages were available elsewhere.”).

137 Nat’l Soc’y of Prof’l Eng’rs v. United States, 435 U.S. 679, 696 (1978).138 See Verizon Commc’ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 407

(2004) (“The opportunity to charge monopoly prices—at least for a short period . . . induces risktaking that produces innovation and economic growth. . . .”).

139 See supra text accompanying notes 82–86.140 The first step could have been stated more broadly and in economic language as requiring

coordinated conduct rather than the more narrow concept of agreement among rivals, but, asdiscussed supra note 83, antitrust law has only limited experience in identifying collusive effectsfrom unilateral conduct or vertical agreements and the modern case law as yet offers little gui-dance on how to do so.


facts suggesting that the exclusionary conduct is anticompetitive, such as ex-cluding all actual or potential rivals. Yet, as I will explain in Part IV, thisdifference also does not undermine the formal parallelism between the trun-cated rules. Rather, the second step operates similarly in both contexts byobviating the need to demonstrate the specific mechanisms by which defend-ants solve the economic problems of making exclusion or collusion work.Those mechanisms are described in the next Part, which identifies the eco-nomic relationship between exclusion and collusion.

III. THE ECONOMICS OF ANTICOMPETITIVE EXCLUSION

The harm to competition that arises from exclusion and collusion can beunderstood within a common economic framework. Indeed, the economic rea-sons for concern about anticompetitive collusion are substantially the same asthe reasons for concern about anticompetitive exclusion.

A. VOLUNTARY AND INVOLUNTARY CARTELS

To see the economic relationship between exclusion and collusion as meansof exercising market power, consider a hypothetical soft drink industry withthree participants: Coke, Pepsi, and Royal Crown (RC). One can imaginethese three rivals reaching an express or tacit (horizontal) agreement to actcollectively as though they were a monopolist, reducing industry output inorder to raise price above the competitive level.141 This outcome could betermed, for reasons that will become clear, a “voluntary” cartel.

Suppose instead that RC does not want to participate in the voluntary cartel.It would prefer to compete rather than to cooperate. In the merger context, onemight describe RC as a “maverick” and be concerned that a merger of RCwith Coke or Pepsi would lead to coordinated competitive effects.142 Moregenerally, if RC would not go along voluntarily with the cartel that Coke andPepsi want to create, then Coke and Pepsi could make it go along by raisingRC’s costs or by making it more difficult for RC to reach customers.143 With

141 This discussion adopts the convention common in the economics literature of describingcompetitive harms in terms of increased prices and a reduction in output. This framework en-compasses a wider range of harms than may be apparent. It includes reductions in product qual-ity, which can be understood as increases in the quality-adjusted price. It may also account forreductions in the rate of innovation, as conduct that reduces competition also tends to discourageinnovation. See generally Jonathan B. Baker, Beyond Schumpeter vs. Arrow: How Antitrust Fos-ters Innovation, 74 ANTITRUST L.J. 575, 579 (2007); Carl Shapiro, Competition and Innovation:Did Arrow Hit the Bull’s Eye?, in THE RATE AND DIRECTION OF INVENTIVE ACTIVITY REVISITED

361 (Josh Lerner & Scott Stern eds., 2012).142 See, e.g., Jonathan B. Baker, Mavericks, Mergers and Exclusion: Proving Coordinated

Competitive Effects Under the Antitrust Laws, 77 N.Y.U. L. REV. 135, 140 (2002).143 If RC is a prospective entrant, Coke may consider a broader range of exclusionary strategies

than would be available if RC is an incumbent firm. In addition to raising RC’s post-entry margi-nal costs of production and distribution, Coke could also deter RC’s entry by making it necessary


higher costs of production or distribution, RC would be forced to cut back itsoutput and raise price, and so permit Coke and Pepsi to reduce their outputand raise their prices without fear that aggressive competition by RC wouldundermine their collusive efforts. The upshot is that all three firms wouldreduce output and raise price, similarly to what would happen if RC wentalong voluntarily with Coke’s and Pepsi’s efforts to collude.144 Because RC iscoerced into participating through the exclusionary conduct of Coke andPepsi, this outcome can be understood as an “involuntary” (or coerced)cartel.145

The “involuntary cartel” terminology is a less natural way of describing theoutcome if Coke is a dominant firm (no Pepsi) and RC is forced to exit ordeterred from entry, as anticompetitive exclusion under such circumstanceswould result in the creation of a literal monopolist.146 Even in this limitingcase, though, the “involuntary cartel” terminology appropriately captures theway the excluding firm forces the excluded rival to do what a cartel partici-pant does voluntarily: avoid aggressive competition. The terminology cap-tures the common adverse economic effect of collusion and exclusion, andfocuses attention on it.

As the soft drink example demonstrates and the “involuntary cartel” termi-nology highlights, exclusion and collusion are complementary methods of ob-taining market power.147 It does not matter to buyers whether the cartel is

for RC to make greater sunk investments in order to enter or by credibly committing to increasethe post-entry competition that RC expects to face. See supra note 75. (The latter strategies canstill be interpreted as raising RC’s marginal costs on the view that a prospective entrant’s margi-nal decision includes whether to enter, not just how much to produce conditional on entry.)

144 The possibility that competition could be harmed through exclusionary conduct has beenwell established in the economics literature for decades. E.g., Richard R. Nelson, IncreasedRents from Increased Costs: A Paradox of Value Theory, 65 J. POL. ECON. 387 (1957); Steven C.Salop & David T. Scheffman, Cost-Raising Strategies, 36 J. INDUS. ECON. 19 (1987); Oliver E.Williamson, Wage Rates as a Barrier to Entry; The Pennington Case in Perspective, 82 Q.J.ECON. 85 (1968).

145 Alternatively, the collusive anticompetitve effects could be described as direct and the ex-clusionary effects described as indirect. ANTITRUST LAW IN PERSPECTIVE, supra note 25, at45–49.

146 The “involuntary cartel” terminology also may mislead if Coke is a dominant firm to theextent it (incorrectly) suggests that excluding firms must solve “cartel problems” in order forexclusion to succeed in that case.

147 Cf. Aaron Edlin & Joseph Farrell, Freedom to Trade and the Competitive Process (Nat’lBureau of Econ. Research, Working Paper No. 16818, 2011), available at http://ssrn.com/abstract=1761581 (collusion and exclusion both harm competition by hindering the process throughwhich buyers and sellers undertake potentially beneficial trades and thereby form improvingcoalitions). But see Thomas G. Krattenmaker, Robert H. Lande & Steven C. Salop, MonopolyPower and Market Power in Antitrust Law, 76 GEO. L.J. 241, 249 (1987) (distinguishing“Stiglerian” (collusive) market power and “Bainian” (exclusionary) market power).


voluntary or involuntary; either way, the same firms collectively reduce out-put and the price that buyers pay increases.148

Exclusion and collusion are also closely related in a second way: they areoften and naturally combined by firms exercising market power.149 Colludingfirms may need to exclude in order for their collusive arrangement to suc-ceed.150 They may find it necessary to deter a cheating member through exclu-sionary conduct, or to exclude fringe rivals or new entrants in order to preventnew competition from undermining their collusive arrangement.151 For exam-ple, the European Commission found that a citric acid cartel threatened dump-ing actions against Chinese firms not participating in their collusivearrangement and targeted price cutting at the customers of the Chinese pro-ducers to deter those rivals from undermining the cartel.152 Indeed, a recentstudy of multiple cartels found that many “use[ ] exclusionary behavior oftenfeatured in monopolization cases to ensure the effectiveness of [their] effortsto restrict output.”153 Similarly, excluding firms may need to collude in orderto exclude successfully154 or to profit collectively from exclusionaryconduct.155

148 In a homogeneous product market, moreover, the exercise of market power reduces industryoutput, regardless of whether the practice is collusive or exclusionary.

149 This discussion illustrates the way that exclusion and collusion permit the firms participat-ing in a market to exercise market power within that market. It does not address the potentialcompetitive consequences of monopoly leveraging (the exploitation of market power in one mar-ket to create market power in another market), except insofar as entry into a complementarymarket would facilitate entry into the market served by the excluding or colluding firms, and theexcluding or colluding firms can maintain their market power in the primary market by foreclos-ing entry by new competitors seeking to sell the complementary product.

150 See ANTITRUST LAW IN PERSPECTIVE, supra note 25, at 235–47 (colluding firms must solvethree “cartel problems,” which include preventing new competition, for their arrangement tosucceed).

151 See JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775, 778 (7th Cir. 1999) (Pos-ner, C.J.) (“JTC, a maverick, was a threat to the cartel—but only if it could find a source ofsupply . . . .”); see generally Baker, supra note 142, at 188–97. Exclusionary conduct may benecessary for coordination among rivals to succeed, regardless of whether the coordination itselfcan be challenged as an illegal agreement.

152 Case COMP/E-1/36.604—Citric Acid, Comm’n Decision, 2002 O.J. (L 239) 18, ¶¶ 116,119, 166.

153 Randal D. Heeb, William E. Kovacic, Robert C. Marshall & Leslie M. Marx, Cartels asTwo-Stage Mechanisms: Implications for the Analysis of Dominant-Firm Conduct, 10 CHI. J.INT’L L. 213, 217 (2009).

154 See Nw. Wholesale Stationers, Inc. v. Pac. Stationery & Printing Co., 472 U.S. 284, 290–91(1985) (alleging agreement among rivals to exclude a competitor); see also Hemphill & Wu,supra note 53.

155 See Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 226–27 (1993)(rejecting predatory pricing claim in part because facts did not support oligopoly recoupmenttheory).


B. EXCLUSION AND ECONOMIC GROWTH

The parallel between voluntary and involuntary cartels provides an eco-nomic basis for treating exclusion and collusion as comparably serious anti-trust offenses. Indeed, anticompetitive exclusion may be the more importantproblem because of the particular threat exclusion poses to economic growth.

When antitrust cases address the suppression of new technologies, prod-ucts, or business models, the disputes are almost always framed as exclusion-ary conduct allegations.156 For example, Microsoft was found to have harmedcompetition in personal computer operating systems by impeding the develop-ment of a new method by which applications software could access operatingsystems, involving the combination of Netscape’s browser and Sun’s Javaprogramming language.157 The D.C. Circuit explained that “it would be inimi-cal to the purpose of the Sherman Act to allow monopolists free rein to squashnascent, albeit unproven, competitors at will—particularly in industriesmarked by rapid technological advance and frequent paradigm shifts.”158

Similarly, much of the relief accepted by the Justice Department and theFederal Communications Commission in their concurrent reviews of Com-cast’s acquisition of NBC Universal programming aimed to protect the devel-opment of nascent competition from a new technology, online videodistribution, and new business models that could threaten Comcast’s marketpower in cable television.159 Exclusionary conduct inhibiting price competi-tion may also harm innovation competition in the same market, as with the

156 See Tim Wu, Taking Innovation Seriously: Antitrust Enforcement if Innovation MatteredMost, 78 ANTITRUST L.J. 313, 316 (2012) (“[I]f innovation mattered most . . . . [antitrust] en-forcement would be primarily concerned with the exclusion of competitors.”). But see UnitedStates v. Automobile Mfrs. Ass’n, 307 F. Supp. 617, 620–21 (C.D. Cal. 1969) (discussing con-sent decree settling allegations of conspiracy to suppress automotive pollution control researchand development); Michelle Goeree & Eric Helland, Do Research Joint Ventures Serve a Collu-sive Function? 5–6 (Institute for Empirical Research in Economics Working Paper No. 1424-059, 2012), available at http://www.iew.uzh.ch/wp/iewwp448.pdf (providing empirical evidencethat research joint ventures among rivals may facilitate collusion). For this purpose, cases alleg-ing conspiracies to exclude firms adopting new technologies, as through group boycott or preda-tory pricing, are counted as exclusionary. See, e.g., Fair Allocation System, Inc., 63 Fed. Reg.43,182 (FTC Aug. 12, 1998) (consent order settling charges that automobile dealers conspired toinduce auto manufacturer to foreclosure rival dealer marketing on the Internet).

157 United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).158 Id. at 79.159 Applications of Comcast Corporation, General Electric Company, and NBC Universal, Inc.,

For Consent to Assign Licenses and Transfer Control of Licensees, Memorandum Opinion andOrder, 26 FCC Rcd. 48 (2011), available at transition.fcc.gov/FCC-11-4.pdf; see CompetitiveImpact Statement, United States v. Comcast Corp., No. 1:11-cv-00106 (D.D.C. Jan. 18, 2011),available at http://www.justice.gov/atr/cases/f266100/266158.pdf. The transaction took the formof a joint venture between Comcast and the previous owner of the programming, General Elec-tric, but was treated as an acquisition because Comcast controlled the joint venture and had theoption to buy out General Electric.


“exclusionary rules” adopted by MasterCard and Visa to prevent memberbanks from issuing American Express or Discover Cards.160

The anticompetitive exclusion of new technologies is not just a modernproblem. Six decades ago, the newspaper monopolist in Lorain Journal im-peded the entry of a rival using a new technology, radio.161 Had the newspapersucceeded, and other newspapers followed suit,162 it is easy to imagine thatfew radio stations in regions with a dominant newspaper would have suc-ceeded unless they were owned by the newspaper, slowing the growth of theradio industry.

These prominent examples make clear that antitrust is an “inclusive” eco-nomic institution that supports economic growth and prosperity by preventingsuccessful incumbent firms and industries from erecting barriers to the entryof rivals with lower costs, superior production technologies, or better prod-ucts.163 The main innovation-related argument otherwise does not question thebenefits of economic growth, whether in individual industries164 or to society

160 United States v. Visa U.S.A., Inc., 344 F.3d 229, 241 (2d Cir. 2003) (upholding districtcourt findings that exclusionary conduct stunted price competition and denied consumer accessto products with new features, and that absent the exclusionary conduct, price competition andinnovation in services would be enhanced).

161 Lorain Journal Co. v. United States, 342 U.S. 143 (1951).162 See Kansas City Star v. United States, 240 F.2d 643 (1957) (news and advertising monopo-

list owned multiple newspapers and radio and television broadcasting stations in the Kansas Cityregion).

163 See DARON ACEMOGLU & JAMES A. ROBINSON, WHY NATIONS FAIL 38–40 (2012) (con-trasting the growth-promoting economic institutions in the United States with the growth-inhib-iting ones in Mexico by comparing Bill Gates, whose technologically innovative company wasprevented from abusing its monopoly by U.S. antitrust enforcers, with Carlos Slim, whose com-pany was conferred monopoly power and protected from competition by Mexican governmentinstitutions). Acemoglu and Robinson attribute economic growth and prosperity primarily to “in-clusive” economic institutions that facilitate entry, investment, and innovation and permit lessefficient firms to be replaced by more efficient ones, id. at 75–79, as opposed to “extractive”economic institutions that “expropriate the resources of the many, erect entry barriers, and sup-press the functioning of markets so that only a few benefit.” Id. at 81. In their view, inclusiveeconomic institutions are typically supported by “inclusive” political institutions that vest powerin a broad coalition or plurality of political groups rather than in a narrow elite. Id. at 80–81,86–87; see also STEPHEN L. PARENTE & EDWARD C. PRESCOTT, BARRIERS TO RICHES 1–34(2000) (attributing differences in living standards across nations importantly to competition-re-ducing policies within less developed countries, put into place to protect the interests of groupsthat benefit from current ways of production, that prevent firms from adopting better productionmethods); cf. Edward L. Glaeser, The Political Risks of Fighting Market Failures: Subversion,Populism and the Government Sponsored Enterprises 8 (Nat’l Bureau of Econ. Research, Work-ing Paper No. 18112, 2012) (“If bargaining across firms is difficult, especially when trying toarrange for large bribes, then competition will lead to less corruption risk than monopoly.”).

164 At the level of individual industries, studies find substantial social gains from new productintroductions. For example, the welfare gain from the introduction of personal computers hasbeen estimated to equal 2–3 percent of consumption expenditure. Jeremy Greenwood & KarenA. Kopecky, Measuring the Welfare Gain from Personal Computers, 51 ECON. INQUIRY 336,336 (2013). Compare Jerry A. Hausman, Valuation of New Goods Under Perfect and ImperfectCompetition, in 58 NBER STUDIES IN INCOME AND WEALTH: THE ECONOMICS OF NEW GOODS


as a whole;165 both skeptics and advocates of antitrust intervention in exclu-sionary conduct settings such as monopolization are concerned with the im-pact of competition policy on growth.166

Instead, those concerned about antitrust enforcement against exclusionaryconduct argue that it could discourage innovation by making it less profitable.Their economic point is that a greater prospect of post-innovation competitioncould reduce the return to innovation.167 But as an argument against antitrustenforcement, it is incomplete because it does not recognize the importance ofcompetitive forces—both pre-innovation product market competition andcompetition in innovation itself—for fostering innovation and economicgrowth.168 Empirical evidence suggests that the latter forces are more impor-

207, 207–35 (Timothy F. Bresnahan & Robert J. Gordon eds., 1996) (estimating substantialsocial welfare gains from the introduction of a differentiated consumer product), with Timothy F.Bresnahan, The Apple-Cinnamon Cheerios War: Valuing New Goods, Identifying MarketPower, and Economic Measurement (undated unpublished manuscript), http://www.stanford.edu/~tbres/Unpublished_Papers/hausman%20recomment.pdf (questioning Hausman’s methodologyand conclusion, but not doubting the likelihood of substantial benefits from new products inhigh-technology sectors). Moreover, other studies find that the social return to innovation sub-stantially exceeds the private return. See Edwin Mansfield, Microeconomics of TechnologicalInnovation, in THE POSITIVE SUM STRATEGY: HARNESSING TECHNOLOGY FOR ECONOMIC

GROWTH 307–11 (Ralph Landau & Nathan Rosenberg eds., 1986); Jeffrey M. Bernstein & M.Ishaq Nadiri, Interindustry R&D Spillovers, Rates of Return, and Production in High-Tech In-dustries, 78 AM. ECON. REV. 429 (1988); see also Zvi Griliches, The Search for R&D Spillovers,94 SCANDINAVIAN J. ECON. S29 (1992); Charles I. Jones & John C. Williams, Measuring theSocial Return to R&D, 113 Q.J. ECON. 1119 (1998).

165 See generally J. Bradford DeLong, Cornucopia: The Pace of Economic Growth in theTwentieth Century (Nat’l Bureau of Econ. Research Working Paper No. 7602, 2000) (highlight-ing the benefits of economy-wide increases in material wealth and productivity).

166 Compare David S. Evans & Keith N. Hylton, The Lawful Acquisition and Exercise of Mo-nopoly Power and Its Implications for the Objectives of Antitrust, 4 COMPETITION POL’Y INT’L

203, 203 (2008) (arguing that antitrust pays excessive attention to the static harms of monopolypricing and insufficient attention to the dynamic benefits of dominant firm innovation), withJonathan B. Baker, “Dynamic Competition” Does Not Excuse Monopolization, 4 COMPETITION

POL’Y INT’L 243, 243–45 (2008) (describing innovation benefits of antitrust enforcement againstmonopolization).

167 For example, in the model analyzed by Hylton and Lin, antitrust enforcement against theexclusionary conduct of dominant firms benefits society by lowering post-innovation consumerprices, but harms society by discouraging innovation. See Keith N. Hylton & Haizhen Lin, Opti-mal Antitrust Enforcement, Dynamic Competition, and Changing Economic Conditions, 77 ANTI-

TRUST L.J. 247, 255 (2010). The model does not incorporate the dynamic benefits of pre-innovation competition in providing an incentive to innovate. See generally Verizon Commc’nsInc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 407 (2004) (the prospect of monop-oly induces risk-taking and innovation).

168 See generally Baker, supra note 141, at 579; Shapiro, supra note 141; cf. TIM WU, THE

MASTER SWITCH: THE RISE AND FALL OF INFORMATION EMPIRES 308 (2010) (“To grant anydominant industrial actor the protection of the state, for whatever reason, is to arrest theSchumpeterian dynamic by which innovation leads to growth, an outcome that is ultimatelynever in the public interest.”); Wu, supra note 156, at 320 (“[A]n innovation-centered antitrustpolicy must make scrutiny of exclusion of innovators its primary concern and a focus ofresources.”).


tant on average.169 As an argument against antitrust, the observation also doesnot recognize the way antitrust enforcement can target industry settings andcategories of behavior where such enforcement can promote innovation.170

Those settings include antitrust enforcement to foster product market competi-tion in “winner-take-all” or “winner-take-most” industries, industries wherethe extent of future competition will be determined mainly by developmentsin technology or regulation, and rapidly growing industries171—all featuresthat frequently characterize high technology sectors.172 In short, antitrust en-forcement against exclusionary conduct is important because it fosters eco-nomic growth and prosperity, not just because it addresses harms to pricecompetition similar to those attacked by enforcement against collusiveconduct.

IV. DIFFERENCES BETWEEN EXCLUSION AND COLLUSION

Notwithstanding the broad doctrinal and economic parallels between volun-tary and involuntary cartels, anticompetitive collusion and exclusion arisethrough different mechanisms. This section explains why those differences donot mean that that antitrust enforcers or courts should downplay exclusionrelative to collusion. It also shows that the truncated legal rules for exclusionand collusion operate similarly: by making it unnecessary to demonstratesome or all elements of the relevant mechanisms.

A. THE ECONOMICS OF “EXCLUSION PROBLEMS”

For an exclusionary strategy to succeed, and thus for the excluding firmssuccessfully to create an involuntary cartel, the excluding firms must solvethree problems: identifying a practical method of exclusion, excluding rivalssufficient to ensure that competition is harmed, and ensuring profitability of

169 See Baker, supra note 141, at 583–87; Shapiro, supra note 141.170 See Baker, supra note 141, at 589; see also Baker, supra note 166 (describing innovation

benefits of antitrust enforcement against monopolization).171 See Baker, supra note 141, at 593–98. Policies increasing pre-innovation competition in

these industries—such as monopolization cases or other antitrust enforcement actions— are un-likely to make much difference to the reward to successful innovation—but can increase pre-innovation competition in both product markets and in innovation, and thus increase overallincentives to innovate.

172 Anticompetitive conduct, whether exclusionary or collusive, can occur in rapidly-innovat-ing high-technology industries, as illustrated by recent government enforcement efforts involvinginformation technology providers. See, e.g., United States v. Microsoft Corp., 253 F.3d 34 (D.C.Cir. 2001) (exclusion); United States v. Adobe Sys., Inc., No. 1:10-cv-01629 (D.D.C. Mar. 18,2011) (final judgment), available at http://www.justice.gov/atr/cases/f272300/272393.htm (col-lusion); United States v. Hynix Semiconductor, Inc., No. CR 05-249 PJH DRAM (N.D. Cal.May 11, 2005) (plea agreement), available at http://www.justice.gov/atr/cases/f209200/209231.pdf (collusion); Intel Corp., FTC Docket No. 9341 (Aug. 4, 2010) (decision and order), availableat http://www.ftc.gov/os/adjpro/d9341/100804inteldo.pdf (exclusion).


their exclusionary strategy.173 These three problems—method, sufficiency,and profitability—may be termed “exclusion problems”174 by analogy to the“cartel problems” that colluding firms must solve in order for a coordinatedarrangement to succeed.175

First, the excluding firms must be able to identify a method of partially orfully excluding some or all rivals.176 The possible methods include four eco-nomic mechanisms for raising rivals’ marginal input costs described byProfessors Krattenmaker and Salop.177 First, excluding firms may create a“bottleneck.”178 The excluding firms can do so by purchasing exclusionaryrights from a sufficient number of the lowest cost suppliers to force excludedrivals to shift to higher cost suppliers or less efficient inputs. The cost increaseleads the excluded rivals to compete less aggressively with the excludingfirms. Second, excluding firms may engage in “real foreclosure.”179 Under thismethod, excluding firms purchase exclusionary rights over a substantial frac-tion of the supply of a key input, and, by withholding that supply, drive up themarket price for the remainder of the input still available to excluded rivals.Again, higher costs would lead the excluded rivals to compete less aggres-

173 These exclusion problems were identified by Professors Krattenmaker and Salop in theirseminal survey article on exclusionary conduct in antitrust. See Krattenmaker & Salop, supranote 9, at 230–49. In that article, Krattenmaker and Salop refer to the first exclusion problem(method) as “raising rivals’ costs,” id. at 230, to the second problem (sufficiency) as “gainingpower over price,” id. at 242, and to the third exclusion problem as “profitability,” using thesame term employed here, id. at 266. The three exclusion problems set forth in the text alsogeneralize the three conditions described as necessary for successful exclusion through verticalagreement set forth in Baker, supra note 74, at 524, which explains that the benefits of thestrategy to the firms undertaking it must exceed its costs (profitability), the excluding firms mustnot cheat on each other (an aspect of sufficiency), and the excluded firm must be unable to avoidthe strategy (method).

174 When there are multiple excluding firms, solving the first and second “exclusion problems”could require coordination among rivals, further illustrating the close connection between collu-sion and exclusion. See Hemphill & Wu, supra note 53 (discussing the ways excluding firmssolve their “cartel problems”).

175 As is well known, colluding firms must find a way to solve three “cartel problems”: reach-ing consensus on terms of coordination, deterring cheating on those terms, and preventing newcompetition. See ANTITRUST LAW IN PERSPECTIVE, supra note 25, at 235–47. These problemsemerge from the Stiglerian deterrence perspective on coordination (as refined by the economicliterature on oligopoly supergames). See Baker, supra note 25, at 149–69 (describing economicsof coordination). They are not tied to the broader perspective on coordination adopted by the2010 Horizontal Merger Guidelines. The Guidelines go beyond the Stiglerian perspective by alsorecognizing as coordination “parallel accommodating conduct” (high price outcomes that resultfrom firm conduct that softens competition when firm strategies do not depend on history andrivals have accommodating reactions). See U.S. Dep’t of Justice & Fed. Trade Comm’n, Hori-zontal Merger Guidelines § 7 (2010), available at http://www.justice.gov/atr/public/guidelines/hmg-2010.pdf.

176 Methods arising in the case law are surveyed informally supra Part II.A.177 See Krattenmaker & Salop, supra note 9, at 234–40.178 Id. at 234.179 Id. at 236.


sively. Third, the excluding firms may act as a “Cartel Ringmaster” by induc-ing multiple suppliers of a key input to sell to the excluded rivals only ondisadvantageous terms, thereby reducing competition from those rivals.180

Fourth, the excluding firms may create a “Frankenstein Monster.”181 The ex-cluding firms would do so by purchasing exclusionary rights from a numberof suppliers of the key input, thereby increasing the likelihood that the re-maining suppliers would successfully collude, expressly or tacitly, to raiseprice to the excluded rivals. With higher input prices, the excluded rivalswould once again be led to compete less aggressively.

The economic mechanisms by which excluding firms foreclose their rivalscould work through raising input prices (input foreclosure), as with the fourmethods just set forth, but they could also operate by reducing rivals’ accessto the market (customer foreclosure). For example, if the rivals benefit fromscale economies and the excluding firms adopt methods that foreclose theexcluded rivals from access to low cost distribution, the excluding firms mayraise their rivals’ costs by reducing their rivals’ scale.182 Indeed, any economicmechanism available for input foreclosure is potentially available for cus-tomer foreclosure, and vice versa.183

If the excluded firms can inexpensively adopt counterstrategies to avoid orevade the exclusionary conduct, the excluding firms will be unable to solvethe first exclusion problem, identifying an exclusionary method.184 In the hy-pothetical soft drink example sketched in Part III.A, if Coke and Pepsi attemptto exclude RC by denying it access to bottlers, but RC can instead obtaincomparable distribution through beer distributors at little cost penalty,185 theexclusionary strategy would not be successful.186 Moreover, if the method of

180 Id. at 238.181 Id. at 240.182 If the excluded rivals must produce and sell at a reduced scale, they may have higher margi-

nal costs. If the excluded rivals can no longer achieve a minimum viable scale, those rivalswould be forced to exit.

183 Some conceptual gymnastics may be required to see the parallel. Consider, for example, anindustry with firms at three levels: input supply, manufacturing, and distribution. Distributionmay more naturally be seen as downstream of manufacturing, but it would not be inappropriateto view it alternatively as a service purchased by the manufacturer. Hence conduct excluding arival from access to distribution could be viewed as input foreclosure as well as customer fore-closure. See also supra note 54.

184 Krattenmaker and Salop analyze rivals’ counterstrategies solely as a profitability issue; herethat issue is also treated as an aspect of the first exclusion problem.

185 This counterstrategy would be unlikely to be available, however, unless beer distributorscan produce bottled products from concentrate.

186 Although customers may have an incentive to help the excluded firms avoid foreclosure,see David T. Scheffman & Pablo T. Spiller, Buyers’ Strategies, Entry Barriers, and Competition,30 ECON. INQUIRY 418 (1992), they need not be able or willing to do so. Customers’ ability toassist excluded firms in executing a counterstrategy is unlikely to be greater than their ability toundermine a voluntary cartel by sponsoring entry. In a market with many buyers, for example, no


exclusion requires coordination between Coke and Pepsi, the inquiry intomethod of exclusion would also include asking whether those firms couldsuccessfully coordinate.187

Second, the exclusionary conduct must be sufficient to harm competition.This condition requires in part that the excluded firm matter competitively;188

its exclusion must relax a competitive constraint on the excluding firms.189 Inaddition, it requires that any remaining competition—whether from rivals notexcluded or not fully excluded, from entrants, or from among the excludingfirms themselves—not undermine what the relaxation of a competitive con-straint has achieved for the excluding firms: their ability to raise marketprices. Accordingly, the excluding firms must prevent their involuntary cartelfrom being undermined through repositioning or output expansion by unex-cluded rivals, by the entry of new competitors, or by cheating among the ex-cluding firms. In a prospective exclusion case, the ability of excluding firmsto solve the sufficiency problem might be inferred from an analysis of marketstructure,190 or from past history of successful exclusion. In a retrospective

individual customer may have sufficient incentive to sponsor entry: doing so would be costly andeach buyer would recognize that most of the benefits would accrue to other buyers rather thanitself.

187 With multiple excluding firms, the excluding firms may have difficulty committing to theirexclusionary method, as they may be unable to avoid the temptation to cheat on the involuntarycartel they have created by foreclosing their excluded rivals. The analysis of whether the exclud-ing firms can successfully overcome this problem, an aspect of sufficiency, would be similar tothe analysis of whether the excluding firms would cheat on the collusive arrangement that wouldhave formed had the excluded firms joined the excluding firms voluntarily. This issue would notarise if there is only a single excluding firm, as would be the case if Coke was a dominant firm.

188 See Joshua D. Wright, Moving Beyond Naıve Foreclosure Analysis, 19 GEO. MASON L.REV. 1163, 1186–87 (2012) (recommending that the share foreclosed by the adoption of anexclusionary rights agreement be measured relative to the share foreclosed but for the adoptionof the agreement).

189 An excluded firm can constrain the excluding firms competitively even if it is not as effi-cient as the excluding firms. For example, if the industry price is $18, the excluding firms havemarginal costs of $10, and the sole excluded firm has a marginal cost of $15, competition wouldbe harmed if the excluding firms raise the price to $20 (say) through foreclosure of the excludedfirm, even though the excluded firm has a higher marginal cost than the excluding firms. In thisexample, the exclusionary conduct would both raise price to consumers and increase the alloca-tive efficiency loss that arises when price exceeds marginal cost. If foreclosure of an inefficientfirm allows a lower cost firm to expand output in its place, as may or may not occur, the resultingproduction cost savings would create a countervailing benefit to aggregate welfare (although thatbenefit would not be cognizable if the welfare criterion looks solely to consumers).

190 If the exclusionary conduct is undertaken by a dominant firm, for example, and the domi-nant firm excludes all significant fringe rivals (those that are not capacity-constrained or other-wise have a high cost of expansion) and entrants, the dominant firm would not face anycompetitive threats. This simple economic idea underlies the truncated legal rule governing ex-clusionary conduct discussed in Part II.B. The second and third open questions about the rulegoverning truncated condemnation of exclusionary conduct suggest some ways of making thisinference other than analyzing the consequences of excluding firm conduct for each actual andpotential rival individually. Supra notes 121–128 and accompanying text.


exclusion case, where market definition may be more difficult,191 evidence ofactual competitive effects may also be available.192

Finally, the exclusionary conduct must be profitable for each excludingfirm.193 Each must reasonably expect that the additional profits it will obtainor maintain through the successful operation of an involuntary cartel wouldexceed the costs it incurs in achieving that arrangement.194 The costs mightinclude, for example, payments to sellers of complements that agree to ex-clude rivals, forgone revenues from reducing price below what the excludingfirms might otherwise charge (e.g., if predatory pricing is alleged as the exclu-sionary mechanism), or forgone profits on lost sales (e.g., if the excludingfirms refuse to deal with buyers that deal with a rival).195 In a retrospectiveexclusion case, profitability might be inferred from observing higher prices orother harms to competition.

The costs of exclusion, and thus the profitability of anticompetitive exclu-sionary conduct, depend upon the nature and scope of the method used toexclude.196 Exclusionary strategies need not be expensive.197 A dominantfirm’s unilateral refusal to deal with suppliers that also supply an entrant orfringe rival, for example, may not be costly if few or no suppliers defect to

191 See generally Baker, supra note 101, at 169–73 (discussing problems that arise in definingmarkets in exclusion settings).

192 This possibility underlies the fourth open question about the truncated legal rule governingexclusionary conduct. See supra notes 129–131 and accompanying text.

193 Krattenmaker and Salop’s notion of profitability includes an evaluation of efficiency justifi-cations. See Krattenmaker & Salop, supra note 9, at 277–82. Efficiencies potentially affect prof-itability to the extent ancillary efficiencies reduce the costs and increase the benefits ofexclusion. Efficiencies are also possible means by which the excluding firms would justify other-wise harmful conduct. Hence, the truncated rule governing exclusionary conduct set forth in PartIV.B requires the absence of a plausible efficiency justification, and efficiencies are consideredas part of a comprehensive reasonableness analysis (as would be undertaken if the truncated ruledoes not apply).

194 When predatory pricing is the exclusionary instrument, this problem is termed “recoup-ment” because the excluding firms bear the cost of exclusion before they earn the rewards. Theprospects for profitability may be challenging to demonstrate in a case that is brought after theexcluding firms have incurred costs of exclusion but before the profits they may earn can beobserved, as may occur with predatory pricing. But profitability may be easier to evaluate whenthe profits from exclusion arise coincident with the costs, as in many non-price exclusionsettings.

195 The costs may also include any expense associated with solving “cartel problems” if multi-ple excluding firms must coordinate in order to exclude rivals or raise price once that exclusionhas occurred. Ancillary benefits to the excluding firms of pursuing exclusionary conduct, such asefficiencies, could reduce the net costs of implementing the method of exclusion.

196 When exclusion is coordinated, moreover, the costs and benefits of exclusion may differamong the excluding firms. See Baker, supra note 46, at 601–02 (noting that one firm may haveborne a disproportionate fraction of the costs and earned the bulk of the benefits of an allegedpredatory pricing conspiracy among cigarette manufacturers).

197 See generally Creighton et al., supra note 81.


dealing with the rival.198 Even when it is expensive for the excluding firms toforeclose their excluded rivals, moreover, the prospective monopoly profitsthe excluding firms obtain from successfully obtaining or protecting marketpower may be great enough to make the expenditure worthwhile.199

B. PROFITABILITY OF PURCHASING AN EXCLUSIONARY RIGHT

The recent economic literature on exclusionary conduct has paid particularattention to identifying conditions under which one exclusionary method, thepurchase of an exclusionary right, would be profitable for the excludingfirms200—thus explaining how the excluding firms solve the third “exclusionproblem” using this strategy.201 A firm purchases an exclusionary right whenit pays a supplier or distributor (or other seller of complements) not to dealwith one or more of the excluding firm’s rivals.202 The arrangement could beexpress, as with Alcoa’s contracts with hydroelectric power producers to pre-vent the generators from supplying electricity to other aluminum manufactur-ers,203 or Microsoft’s agreements with Internet access providers to limit

198 See, e.g., Lorain Journal Co. v. United States, 342 U.S. 143 (1951).199 See generally Krattenmaker & Salop, supra note 9, at 273–77 (noting that rivals benefiting

from an involuntary cartel may find it necessary to share their monopoly profits with sophisti-cated input suppliers).

200 The analysis of the profitability of this exclusionary strategy is complicated by the need toaccount for the interaction between the excluding firms and the vertically related firms (or othersellers of complements) that agree not to deal with the excluded firms or otherwise foreclosethem.

201 Other exclusionary methods, not discussed in detail here, may resemble the purchase of anexclusionary right because they may also involve contracts or understandings between the ex-cluding firm and sellers of complements. These may include “most favored nations” (MFN, or“most favored customer”) clauses, which can be employed by dominant firms to ensure thatfringe rivals and entrants cannot lower their costs by obtaining lower prices from sellers ofcomplements, bundled (or loyalty) discounts (or rebates) offered by manufacturers to dealers,slotting allowances, and resale price maintenance. To the extent these methods operate like thepurchase of an exclusionary right, the economic analysis in the text is likely relevant to under-standing their competitive implications. See John Asker & Heski Bar-Isaac, Vertical PracticesFacilitating Exclusion (NYU Working Paper, 2012) available at http://ssrn.com/abstract=218182(a manufacturer can employ a range of vertical practices to share profits from its exercise ofmarket power with its dealers, thereby giving the dealers an incentive not to deal with an entrantinto manufacturing); cf. Timothy J. Brennan, Getting Exclusion Cases Right: Intel and Beyond,CPI ANTITRUST CHRONICLE 5–7, Dec. 2011 (1), available at https://www.competitionpolicyinternational.com/dec-11 (exclusionary conduct that operates by suppressing competition in the mar-ket for a complementary product can profit firms selling complements). (In other settings, thoughthese practices may harm competition through means other than exclusion (including facilitatingcollusion, dampening competition, or facilitating anticompetitive price discrimination), or theymay permit firms to achieve efficiencies.)

202 Although the discussion below will be framed by an example in which the excluded firmwas cut off from the supply of a key input, the same analysis would apply if the excluded firmwas cut off from an important distribution channel.

203 See United States v. Aluminum Co. of Am., 148 F.2d 416, 422–23 (2d Cir. 1945) (Alcoa)(describing agreements addressed in a 1912 government antitrust enforcement action).


distribution of Netscape’s browser.204 The arrangement could also be tacit, aswhen a manufacturer pays a generous wholesale price to a supplier and thesupplier returns the favor by not dealing with one or more rivalmanufacturers.205

Several factors affect the profitability of exclusion through the purchase ofan exclusionary right.206 Some of these factors were addressed by the SeventhCircuit in JTC Petroleum Co. v. Piasa Motor Fuels, Inc.207 In that case, thecourt evaluated the plausibility of an alleged bid-rigging scheme in which thedefendant producers were said to have protected their collusive arrangementby inducing suppliers of a key input to refuse to sell that input to JTC, amaverick rival that otherwise would have cheated on the cartel. The appellatecourt overruled the district court’s award of summary judgment to defendants,holding that a rational jury could conclude that the excluded firm was thevictim of a supplier boycott organized by its rivals.

The four factors set forth below presume that a dominant firm is engaged inthe exclusionary conduct, without addressing the additional complications thatarise when firms coordinate to purchase exclusionary rights, as was alleged inJTC Petroleum. Instead, the discussion will suppose that one of the allegedexcluding firms in the case, Piasa, paid asphalt suppliers not to supply one ofPiasa’s rival’s, JTC. This arrangement could have been negotiated explicitlyand memorialized contractually, or it could have been informal and tacit, asmight arise if Piasa paid generously for its asphalt purchases while simultane-ously making clear to the supplier the high value it places on supplierloyalty.208

204 United States v. Microsoft Corp., 253 F.3d 34, 67–68 (D.C. Cir. 2001) (en banc).205 See, e.g., JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775, 778 (7th Cir. 1999)

(Posner, C.J.) (noting that cartel members may have induced suppliers to refuse to sell to thecartel’s maverick rival by paying a higher price for the supplied product).

206 See generally MICHAEL D. WHINSTON, LECTURES ON ANTITRUST ECONOMICS 133–97(2006) (surveying the economic literature on exclusionary vertical contracts, with attention toantitrust applications); Claudia M. Landeo, Exclusionary Vertical Restraints and Antitrust: Ex-perimental Law and Economics Contributions, in RESEARCH HANDBOOK ON BEHAVIORAL LAW

AND ECONOMICS (Kathryn Zeiler & Joshua Teitelbaum eds., forthcoming) (surveying the experi-mental economic literature on exclusionary vertical contracts, with attention to antitrustapplications).

207 JTC Petroleum, 190 F.3d 775. The terminology employed here highlights the vertical struc-ture of the exclusionary contracts but may be confusing to a reader familiar with the opinion. Inthe opinion, the firms described here as producers are termed “applicators” (they are road con-tractors), and the firms described here as suppliers are termed “producers” (because they produceasphalt for use by the applicators). Id. at 776.

208 An informal or tacit exclusionary arrangement presents an additional difficulty not presentif the exclusionary right is purchased through a contract. Piasa might wish to subsidize one ormore suppliers as an implicit quid pro quo for exclusivity, perhaps by paying generously for itsown input purchases. But Piasa could not profitably do so if the suppliers would act opportunisti-cally (by taking the payment without following through by cutting off JTC), as the resultingcompetition between Piasa and JTC would lower Piasa’s profits. However, the suppliers would


The first factor is the relative profitability of success for Piasa and JTC. ForPiasa, success would mean excluding JTC and thus obtaining or maintainingmarket power. For JTC, success would mean participating in the market byavoiding exclusion. The relative profitability of success matters because thefirm with the greater financial advantage should its strategy succeed is betterpositioned to win a bidding war over whether the supplier will sell asphalt toJTC. Piasa may have a financial advantage arising from the monopoly profitsit would earn if it successfully excluded JTC,209 while JTC may have a finan-cial advantage if can produce at lower cost than Piasa.210 In the JTC Petroleumcase, the Seventh Circuit addressed this factor by explaining that the cartelprofits could have given the producers a fund with which to compensate theinput suppliers.211

The second factor affecting the profitability to Piasa of purchasing an ex-clusionary right from an input supplier is Piasa’s ability to limit the scope andcost of its investments in exclusion through careful targeting.212 For example,if Piasa must pay the supplier to deal with Piasa exclusively, and thus pay itnot to sell to a wide range of other firms (perhaps including rivals to Piasa andfirms producing in geographic markets other than those in which Piasa andJTC compete), Piasa may have to pay more than it would pay to induce thesupplier not to sell to just one firm, JTC. Piasa was more likely to find the

not necessarily act opportunistically, for example if they fear that Piasa would respond by notpurchasing from them.

209 Piasa would no longer have that advantage if JTC would reasonably expect that if it man-ages to avoid exclusion, Piasa would not compete aggressively against it. See Chiara Fumagalli& Massimo Motta, Exclusive Dealing and Entry, When Buyers Compete, 96 AM. ECON. REV.785 (2006) (noting that if a successful entrant would be likely to take a substantial fraction of themarket from a dominant incumbent, distributors may see greater benefit in dealing with theentrant than they would if the entrant’s prospects were more limited).

210 JTC might have a more efficient production technology or better business model, for exam-ple. In other settings, the excluded firm’s financial advantage should it successfully enter mightcome from its ability to offer an attractive new or improved product. Even if JTC has substantialfinancial resources, moreover, its ability to convince suppliers to sell it asphalt will also dependon the extent to which an individual supplier would expect to share in the profits from JTC’smarketplace success. The more that competition among asphalt suppliers to serve JTC would beexpected to dissipate supplier profits, the less interest a supplier would have in serving JTCrather than accepting Piasa’s request for exclusivity. See WHINSTON, supra note 206, at 148–49;John Simpson & Abraham L. Wickelgren, Naked Exclusion, Efficient Breach, and DownstreamCompetition, 97 AM. ECON. REV. 1305 (2007).

211 JTC Petroleum, 190 F.3d at 778. There was no suggestion that the excluded firm was anefficient producer that would have earned higher profits in a competitive market than the allegedcartelists would have obtained through successful bid-rigging. See id.

212 See Claudia M. Landeo & Kathryn E. Spier, Naked Exclusion: An Experimental Study ofContracts with Externalities, 99 AM. ECON. REV. 1850 (2009) (discrimination by incumbentseller facilitates exclusion); Patrick DeGraba, Naked Exclusion by a Dominant Supplier: Exclu-sive Contracting and Loyalty Discounts (Fed. Trade Comm’n, Working Paper No. 306, 2010),available at http://www.ftc.gov/be/workpapers/wp306.pdf (a dominant input supplier can pre-vent a smaller rival from expanding by using exclusive contracts and price discriminating basedon an end user’s likelihood of purchasing products made with the rival’s input).


strategy of purchasing an exclusionary right to be profitable if it could excludeJTC cheaply.213 This factor was not addressed explicitly in the Seventh Cir-cuit’s JTC Petroleum decision, perhaps because the alleged exclusionary con-duct in that case targeted only the maverick rival and was not so broad inscope as to call into question the profitability of this exclusionary strategy.

The third factor affecting the profitability of exclusion through purchase ofan exclusionary right is whether the suppliers expect Piasa to succeed in ex-cluding JTC. This factor would matter most if there were many suppliers (or,equivalently, many distributors), and the excluded rival (JTC) must obtain thekey input from several suppliers (but not necessarily all) in order to succeed.Under such circumstances, each supplier’s decision to contract with Piasa notto deal with JTC—and the price it will demand from Piasa in order to excludeJTC—may depend on each supplier’s expectations about JTC’s likely successin reaching a deal with other suppliers.214 If each supplier thinks that JTC isunlikely to find enough supply to compete successfully by contracting with

213 To similar effect, Alcoa apparently targeted entry by rival aluminum producers by con-tracting with hydroelectric power producers to prevent the generators from supplying electricityto other aluminum manufacturers. The contracts were targeted because they did not preclude thepower producers from selling electricity to firms producing other products. See United States v.Aluminum Co. of Am., 148 F.2d 416, 422–23 (2d Cir. 1945) (Alcoa) (describing agreementsaddressed in a 1912 government antitrust enforcement action). Selective price matching may alsooperate as a targeted exclusionary strategy. For example, a manufacturer may exclude a rival bymatching any price reduction that the rival offers customers with a comparably low price. Thatstrategy would be targeted if the manufacturer only cuts price to those customers solicited specif-ically by the excluded rival. But if the manufacturer cannot commit to limited targeting in re-sponse to competition, and must fight an aggressive rival with an (expensive) across-the-boardprice reduction, then the exclusionary price-cutting strategy may not be profitable. Cf. MICHAEL

E. PORTER, COMPETITIVE ADVANTAGE 500–01, 511 (1985) (recommending “plac[ing] potentialchallengers at a relative cost disadvantage” by “targeting” price cuts on “products that are likelyinitial purchases by new buyers” or by “localizing” the response to rival price cutting “to particu-larly vulnerable buyers” rather than across-the board to reduce the cost of the response); BRUCE

GREENWALD & JUDD KAHN, COMPETITION DEMYSTIFIED 231 (2005) (recommending than anincumbent respond to entry by “punish[ing] the newcomer as severely as possible at the lowestpossible cost to itself”). But see Judith R. Gelman & Steven C. Salop, Judo Economics: CapacityLimitation and Coupon Competition, 14 BELL J. ECON. 315, 316 n.2 (1983) (noting that evensmall sunk expenditures may be sufficient to prevent entry by serving as a credible commitmentto post-entry competition across-the-board). I am grateful to Aaron Edlin for sharing his insightsinto the exclusionary potential of price-matching by incumbent firms.

214 Supplier expectations may depend on the nature of the interdependence among their deci-sions. One supplier’s decision to contract with JTC may confer a positive externality on othersuppliers, as by making it more likely that JTC’s strategy to compete will succeed. For example,each supplier’s agreement to supply JTC may make it more likely that JTC will convince enoughother suppliers to do so, and in consequence, may make it more valuable for each undecidedsupplier to agree to supply JTC. In other settings, however, the decision by one supplier tocontract with JTC would reduce the sales and profits available to other suppliers, conferring anegative externality that lessens other suppliers’ gains from contracting with JTC. Supplier ex-pectations may also depend on whether the excluding firms can improve their odds of success bysupplementing the strategy of purchasing an exclusionary right with additional strategies forraising rivals’ costs or reducing their access to the market.


other suppliers, it would do better by accepting a payment from Piasa forexclusivity—even if it gets little or nothing for doing so—than by contractingwith JTC.215 The critical role of supplier expectations in this dynamic maymean that the greater the number of suppliers JTC needs to contract with forviability, the less likely that any individual supplier would expect JTC to suc-ceed and, thus, the more likely that each supplier would agree with Piasa notto deal with JTC.216

The JTC Petroleum opinion did not address this factor explicitly, presuma-bly because there were only three suppliers, and JTC would likely succeed ifit contracted with any one of them. But the opinion did consider supplierexpectations by providing three possible reasons why no supplier would actagainst the cartel’s wishes: the cartel might be able to coerce the suppliers bythreatening to exercise monopsony power, the cartel’s payments to the suppli-ers may have been too large to resist, and the suppliers themselves may havecolluded (as plaintiffs had alleged), so each supplier’s individual decision notto sell to the maverick producer may also have been supported by the threat ofpunishments from the other suppliers.217 These possibilities suggest ways thatPiasa could have succeeded in excluding JTC by influencing supplier expecta-tions if JTC had needed to assemble multiple suppliers in order to compete.

The final factor affecting the profitability of Piasa’s purchase of an exclu-sionary right is the relative ability of Piasa and JTC to make credible commit-ments to the suppliers. Piasa’s bargaining position with the suppliers would beimproved if it can commit that it will not purchase from suppliers that sell toJTC.218 If Piasa can convince the suppliers that they must choose between itand JTC—that Piasa will follow through on its intention not to buy fromsuppliers that sell to JTC—then the suppliers may prefer to work with Piasaby cutting off JTC even if Piasa offers little or nothing in payment for exclu-sivity. But if the suppliers think that in the event they sell to JTC, Piasa wouldprefer to continue to purchase from them notwithstanding Piasa’s displeasurewith their decision to do business with JTC, then Piasa will have to pay more

215 If JTC needs multiple suppliers to survive but each supplier thinks the other suppliers willaccept an exclusive deal with Piasa, no supplier would break ranks to deal with JTC even if Piasapays nothing. After all, a supplier that expects JTC to fail would not want to ruin its relationshipwith Piasa by contracting with JTC.

216 Although the pure “naked exclusion” model has two equilibria, one in which all bottlersagree to exclusivity with Piasa and one in which none do so, the exclusion equilibrium dominatesif Piasa can convince even a small number of suppliers not to deal with JTC. See Rasmusen etal., supra note 80, at 1143–44; Ilya R. Segal & Michael D. Whinston, Naked Exclusion: Com-ment, 90 AM. ECON. REV. 296 (2000).

217 JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775, 778–79 (7th Cir. 1999) (Pos-ner, C.J.).

218 Cf. Steven C. Salop & R. Craig Romaine, Preserving Monopoly: Economic Analysis, LegalStandards, and Microsoft, 7 GEO. MASON L. REV. 617, 640–42 (1999) (discussing credibility ofpredatory threats by a dominant firm).


to induce the suppliers not to deal with JTC, possibly making it uneconomicfor Piasa to employ this exclusionary strategy.219 The court did not explicitlyconsider this factor, but it implicitly addressed the credibility of cartel threatsnot to purchase from suppliers that sell to JTC and other maverick producersby suggesting the possibility that the colluding producers had coerced thesuppliers.

The three exclusion problems (means, sufficiency, and profitability) differfrom the three cartel problems (reaching consensus, deterring deviation, andpreventing new competition) because the mechanisms by which firms achievean involuntary cartel and a voluntary one are not the same.220 But as will beseen in the next Part, the legal rules governing structured review of allegedexclusionary and collusive conduct limit the factors that a court must considerbefore truncated condemnation in analogous ways—providing additional sup-port for the conclusion that antitrust doctrine treats anticompetitive exclusionand anticompetitive collusion as comparably serious offenses.

C. EXCLUSION PROBLEMS AND TRUNCATED REASONABLENESS REVIEW

The economic logic underlying truncated reasonableness review is clarifiedby recognizing that firms may obtain or maintain market power by solvingexclusion problems or collusion problems. The structured rules that permitcondemnation of exclusionary or collusive conduct, described above in PartII.B, prove harm to competition through limited factual showings.221 In thecollusion context, plaintiff may rely on facial analysis or categorization of theagreement (as price fixing or market division), or on actual effects evidence.222

In the exclusion context, plaintiff may rely, at a minimum, on evidence that allactual or potential rivals other than insignificant competitors have beenexcluded.223

219 On the other hand, if JTC can commit to limiting its output (as by adopting a high-cost orcapacity-constrained production process, or locating its facilities so it can only serve some cus-tomers inexpensively), then Piasa may conclude that it would be less costly to allow JTC to enterrather than spend what would be required to prevent JTC from obtaining key inputs. See Gelman& Salop, supra note 213, at 315.

220 The problems may overlap, however, if colluding firms must solve their exclusion problemsin order to prevent new competition by rivals or entrants, or if multiple excluding firms mustsolve their cartel problems in order to exclude.

221 Both rules have two elements other than harm to competition. Each incorporates a conductpredicate—exclusion in one case, agreement in the other—and each requires the absence of aplausible efficiency justification for the exclusionary conduct or collusive arrangement at issue.

222 See supra notes 84–86 and accompanying text.223 See supra note 97 and accompanying text. Some of the open questions ask whether two

other limited forms of proof, actual effects evidence and excluding firm market power, couldalso be sufficient to demonstrate harm to competition. See supra notes 121–122, 129–131, andaccompanying text.


Those truncated rules limit the detail with which competitive effects areanalyzed. They may make it unnecessary to examine whether a defendant canor did solve each exclusion problem (in an exclusionary effects case) or eachcartel problem (in a collusive effects case), as would be relevant when evalu-ating the conduct under the comprehensive rule of reason.224 The truncatedcondemnation rule for exclusionary conduct infers harm to competition fromevidence that all actual or potential rivals other than insignificant competitorshave been excluded through conduct lacking a plausible efficiency justifica-tion. If these facts can be demonstrated, the conduct can be found to harmcompetition by presuming (or inferring) that firms can solve one of their ex-clusion problems, profitability, without specifically analyzing it. Two of theopen questions—the possibility that proof of excluding-firm market powermay permit the inference that all rivals are or likely would be excluded fromevidence that one has been excluded, and the possibility that harm to competi-tion could be shown through actual effects evidence in a retrospective exclu-sion case—raise the possibility of presuming (or inferring) that firms can alsosolve a second exclusion problem, sufficiency, again without specifically ana-lyzing it.

Whether exclusion or collusion is alleged, truncation of the reasonablenessreview means that adverse competitive effects can be inferred without show-ing fully why the anticompetitive mechanism worked: why it was profitable in

224 See, e.g., United States v. Visa U.S.A., Inc., 344 F.3d 229 (2d Cir. 2003) (explicitly analyz-ing the method of exclusion and its sufficiency, and implicitly analyzing its profitability byrecognizing that it protected excluding-firm market power from erosion). A wide range of otherevidence could be relevant under the comprehensive rule of reason to determining whether de-fendants in an exclusionary conduct case have solved their exclusion problems. For example, theavailability of potential alternative sources of distribution to an excluded manufacturer may bearon whether the plaintiff was substantially or insignificantly excluded, as may the duration andcosts of terminating the exclusivity agreements that limit the excluded firm’s access to custom-ers, and the percentage of the market foreclosed to the excluded firm by the conduct at issue maybe relevant to assessing the sufficiency of the alleged exclusionary acts to create harm to compe-tition. E.g., Omega Envtl., Inc. v. Gilbarco, Inc., 127 F.3d 1157 (9th Cir. 1997); cf. NicSand, Inc.v. 3M Co., 507 F.3d 442 (6th Cir. 2007) (finding that an excluded rival challenging a dominantmanufacturer’s multi-year distribution contracts with all major retailers lacked antitrust injurywhen the excluded rival was formerly dominant, previously had exclusive distribution arrange-ments with most of the leading retailers, and had an equal opportunity to compete for exclusivitywith the new dominant firm). The focus on “coercion” by the excluding firm in a recent appellatedecision can be understood either as an aspect of the inquiry into whether the plaintiff has alter-natives for avoiding exclusion (an inquiry into “method”) or as an aspect of an inquiry intowhether the defendant had a legitimate business justification for the practice. Race Tires Am.,Inc. v. Hoosier Racing Tire Corp., 614 F.3d 57, 77–79 (3d Cir. 2010). Comprehensive reasona-bleness analysis of exclusionary conduct is the same regardless of whether the conduct is chal-lenged by the government or an excluded firm or whether the case is brought under Sherman ActSection 1 (as a vertical agreement), Sherman Act Section 2 (as monopolization or an attempt tomonopolize), Clayton Act Section 3 (as exclusive dealing), or FTC Act Section 5. See generallyHOVENKAMP, supra note 36, § 10.9 (noting that exclusive dealing has been condemned underseveral statutes).


both settings, perhaps why the method was sufficient in an exclusion case, andhow the firms deterred cheating and prevented new competition in a collusioncase. The specific methods by which firms exercise market power differacross the settings, but the burden on plaintiff is reduced in an analogous wayin each. Again the formal structure of the truncated doctrinal rule does nottreat exclusion differently from collusion.

V. POLICY CONSIDERATIONS DO NOT JUSTIFYDOWNPLAYING EXCLUSION

Although the rhetorical consensus for treating exclusion as a lesser offenseis commonly asserted without explicit justification, its defenders sometimesargue that false positives (convictions) are more likely or more costly withexclusionary violations than collusive ones, while false negatives (acquittals)are less likely or less costly with exclusion than collusion.225 This framingadopts an “error cost” perspective to evaluate antitrust rules, under which thebest rule minimizes total social costs.226 This general approach toward evaluat-ing legal rules has been employed by the Supreme Court in recent antitrustdecisions.227

In antitrust applications, the costs to society that need to be considered ex-tend beyond litigation costs and the consequences of alternative decisions tothe parties to a case; they also include the benefits (negative costs) of deter-ring harmful conduct and costs of chilling beneficial conduct, throughout the

225 See, e.g., Keith N. Hylton, The Law and Economics of Monopolization Standards, in ANTI-

TRUST LAW AND ECONOMICS 82, 102 (Keith N. Hylton ed., 2010).226 The relevant social costs are commonly described as the costs of false positives and false

negatives, along with the transaction costs associated with the use of the legal process. See DOJSECTION 2 REPORT (WITHDRAWN), supra note 11, at 15–18 (endorsing error cost framework forthe evaluation of Section 2 standards). Transaction costs include more than the costs of litigation;they also include costs associated with information-gathering by the institution specifying deci-sion rules. See C. Frederick Beckner III & Steven C. Salop, Decision Theory and Antitrust Rules,67 ANTITRUST L.J. 41 (1999).

227 See, e.g., Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 895 (2007);(explaining that per se rules may “increase the total cost of the antitrust system” even when they“decrease administrative costs” if they “prohibit[ ] procompetitive conduct the antitrust lawsshould encourage” or “increase litigation costs by promoting frivolous suits against legitimatepractices” and overruling rule of per se illegality against vertical price restraints); VerizonCommc’ns Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 408 (2004) (describing aneed to be “very cautious” in finding an antitrust violation when a dominant firm unilaterallyrefuses to cooperate with a rival “because of the uncertain virtue of forced information sharingand the difficulty of identifying and remedying anticompetitive conduct by a single firm”); id. at414 (expressing concern with the “cost of false positives” arising from the possibility that“generalist antitrust court” would need to enforce a complex statutory scheme in a dynamicindustry); Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 226 (1993)(stating that in predatory pricing cases, “the costs of an erroneous finding of liability are high”because of the danger that false convictions would chill procompetitive price cutting).


economy.228 But it can be difficult to account for economy-wide effects withinthe error cost framework.229 Moreover, when deterrence and chilling effectsare accounted for, substantive legal rules can properly be compared on thebasis of their error costs only if the comparison holds constant a wide range ofbackground institutions: both antitrust enforcement institutions, such as rulesgoverning burdens of proof,230 which can vary in impact across doctrinal cate-gories,231 and non-antitrust institutions, such as the scope of intellectual prop-erty rights.232 Notwithstanding these conceptual and practical difficulties, thissection will discuss antitrust rules in terms of the familiar categories of falsepositives and false negatives.

Given that the similarity in the economic reasons for concern about an-ticompetitive collusion and anticompetitive exclusion,233 and the similar struc-ture to the legal rules governing exclusion and collusion,234 an error costargument for downplaying exclusion relative to collusion cannot turn on howconduct would be evaluated by an omniscient judge faithfully applying therules. Rather, the two leading and closely related policy arguments offered fordownplaying exclusion relative to collusion rely on arguments about the rela-tive likelihood and magnitude of mistakes in the antitrust review of collusiveconduct compared with exclusionary conduct. The first supposes that it is

228 See, e.g., Brooke Group, 509 U.S. at 223 (permitting predatory pricing enforcement basedon above cost prices would create “intolerable risks of chilling legitimate price-cutting”); AlliedOrthopedic Appliances Inc. v. Tyco Health Care Grp., 592 F.3d 991, 1000 (9th Cir. 2010) (not-ing that courts should not find monopolization when the alleged exclusionary act is a non-shamproduct design improvement because of the danger of dampening technological innovation andthe difficulty of weighing uncertain future benefits against current competitive harms). Deter-rence considerations are particularly important in evaluating antitrust rules. See generallyJonathan B. Baker, The Case for Antitrust Enforcement, 17 J. ECON. PERSP., Autumn 2003, at 27.

229 False positives and false negatives may not neatly map onto overdeterrence and underdeter-rence, respectively, because the deterrence consequences of legal errors depend in part on theway the errors affect the marginal costs and benefits to firms of taking precautions to avoidviolations. See generally Warren F. Schwartz, Legal Error, in ENCYCLOPEDIA OF LAW AND ECO-

NOMICS 1029, 1029–38 (Boudewijn Bouckaert & Garrit De Geest eds., 1999), available at http://encyclo.findlaw.com/0790book.pdf.

230 See generally Louis Kaplow, Burden of Proof, 121 YALE L.J. 738 (2012); Wickelgren,supra note 77, at 54 (noting that the consequences of greater accuracy at trial for firm conductdepend in part on whether firms can predict how trial outcomes will change, whether settlementoutcomes are improved, and how decisions to sue or settle are affected).

231 See Stephen Calkins, Summary Judgment, Motions to Dismiss, and Other Examples ofEquilibrating Tendencies in the Antitrust System, 74 GEO. L.J. 1065, 1127–39 (1986) (docu-menting the way the antitrust treble damages remedy has shaped substantive and proceduralantitrust law across doctrinal categories).

232 The balance of error costs could change if the antitrust rules themselves change, as whenthe Supreme Court modified many legal rules in response to Chicago School concerns that theyimpeded efficiency enhancing business conduct, or if the background institutions change evenwhen the rules do not.

233 See supra Part III.A.234 See supra Part II.B.


harder to tell apart harmful and beneficial conduct when exclusion is alleged,so enforcers and courts are more likely to make errors in that setting.235 Thesecond contends that false positives are more dangerous when exclusion isalleged because they are more likely to chill beneficial conduct like price cut-ting and new product introductions—making it more important that enforcersand courts avoid errors in the exclusion setting.236 If errors are more frequentand more costly to society when exclusionary conduct is alleged, as these twostories claim, enforcers and courts should be more cautious in challengingsuch conduct.

The enforcement agencies do challenge anticompetitive collusion more fre-quently than they challenge anticompetitive exclusion. Since 1980, U.S. casesinvolving horizontal restraints, a collusion-oriented doctrinal category, havebeen brought substantially more often than cases in doctrinal categories whereexclusion is more likely to be found (monopolization and vertical agree-ments).237 Although this observation may seem to follow from agency viewsregarding relative error costs,238 or even to vindicate that perspective, there is

235 See Frank H. Easterbrook, When Is It Worthwhile to Use Courts to Search for ExclusionaryConduct?, 2003 COLUM. BUS. L. REV. 345, 345 (“[C]ompetitive and exclusionary conduct lookalike.”); DOJ SECTION 2 REPORT (WITHDRAWN), supra note 11, at 13 (“[O]ften the same conductcan both generate efficiencies and exclude competitors.”).

236 See Easterbrook, supra note 235, at 347.237 See generally Kovacic, supra note 30, at 377. Including Robinson-Patman violations as an

exclusion category would not alter this conclusion. The frequency of anticompetitive collusionchallenges relative to harmful exclusion allegations in court cases is also likely skewed towardcollusion, even though the frequency of court cases is likely driven by private litigation ratherthan agency enforcement actions. A study of the private treble damages cases filed between 1973and 1983 in five districts found substantially more raised horizontal allegations (which tend toinvolve collusive conduct) than vertical allegations (which more often involve exclusionary con-duct). See Steven C. Salop & Lawrence J. White, Treble Damages Reform: Implications of theGeorgetown Project, 55 ANTITRUST L.J. 73, 74 (1986) (finding that “52.8 percent of cases incor-porated vertical allegations [while] 71.6 percent [incorporated]] horizontal allegations”). Manycases included both horizontal and vertical claims, consistent with the possibility that exclusion-ary conduct was part of a collusive scheme (as can be the case even when the plaintiff is a rival).See, e.g., JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775 (7th Cir. 1999). Thesestatistics were not driven by cases brought by standalone competitors, as they accounted for only22.9 percent of the filings in the sample. See Salop & White, supra, at 74. The explanations andimplications of the disparity in the frequency of the two types of agency cases, discussed below,would also likely apply to the interpretation of the relative frequency of private cases, as privateplaintiffs and their attorneys often face a similar cost-benefit calculus as the enforcement agen-cies in allocating their resources and many private cases are follow-ons to governmentinvestigations.

238 Cf. DOJ SECTION 2 REPORT (WITHDRAWN), supra note 11, at 5, 8 (relying on the U.Sexperience applying Section 2 to derive broad principles). In recent years, moreover, the agen-cies may have begun to prioritize collusion in ways that go beyond allocating enforcement re-sources. The agencies have targeted collusion for increased penalties, greater internationalcooperation, and the increased use of leniency programs to provide an incentive for colludingfirms to come forward. Scott D. Hammond, Deputy Assistant Att’y Gen., U.S. Dep’t of Justice,The Evolution of Criminal Antitrust Enforcement Over the Last Two Decades (Feb. 25, 2010),available at http://www.justice.gov/atr/public/speeches/255515.htm; see Stephen Labaton, The


actually no necessary relationship between the distribution of errors and theircosts and the relative frequency of cases. Even if the allocation of agencycases provides a reliable guide to the relative frequency of the two types ofcompetitive problems, the observed enforcement pattern does not mean thaterrors in resolving antitrust allegations, whether false positives or false nega-tives, are either more frequent or more costly in exclusion settings. Neverthe-less, to evaluate the relative frequency and magnitude of errors, it is useful tofirst examine why the agencies emphasize collusion over exclusion in caseselection.

A. RELATIVE FREQUENCY OF ERRORS

To begin with frequency, the connection between the relative number ofcollusion and exclusion challenges and the frequency of errors depends in parton why the agencies bring more collusion cases than exclusion cases. Thereare a number of possible explanations, including the following four. First, theagencies may bring more collusion cases because the antitrust laws deter an-ticompetitive exclusion more effectively than they deter anticompetitive collu-sion.239 This explanation makes sense: collusion is likely easier to hide thanexclusion; exclusion can often more easily or reliably be achieved throughconduct that would not violate the antitrust laws (such as lobbying for govern-mentally created entry barriers240); and firms with an incentive to avoid anti-

World Gets Tough on Fixing Prices, N.Y. TIMES, June 3, 2001, http://www.nytimes.com/2001/06/03/business/the-world-gets-tough-on-fixing-prices.html. At the same time, they have debatedwhether courts should relax the legal rule governing monopolization, which is almost always anexclusionary offense, in order to raise the practical burden on plaintiffs. See supra note 37. Bycontrast, European competition policymakers tend to express greater concern with exclusionaryconduct. See, e.g., Eleanor Fox & Daniel Crane, GLOBAL ISSUES IN ANTITRUST AND COMPETI-

TION LAW 122, 123–29 (2010) (contrasting the EU test for predatory pricing with the “veryconservative” approach of U.S. courts); id. at 130 (observing that EU law is more likely torequire that a dominant firm deal with its rival than U.S. law); id. at 143 (noting “a significantdivide” between the United States and the European Union on using competition policy to ad-dress margin squeezes by regulated firms); id. at 197 (“The European Union . . . is less permis-sive [than U.S. law on vertical restraints].”).

239 Many examples of anticompetitive collusion, exclusion, and mergers have been docu-mented across a range of industries during periods of lax antitrust enforcement, such as thequarter century after the Sherman Act was enacted, suggesting that the antitrust laws have de-terred a great deal of anticompetitive conduct. See Baker, supra note 228, at 36–38.

240 Cf. James C. Cooper, Paul A. Pautler & Todd J. Zywicki, Theory and Practice of Competi-tion Advocacy at the FTC, 72 ANTITRUST L.J. 1091 (2005) (describing trends in the FTC’s com-petition advocacy program, which questions proposed federal or state legislation and regulationsthat threaten to impede competition). The Federal Trade Commission has also emphasized theimportance of construing the state action exemption to the antitrust laws narrowly in order todiscourage the manipulation of regulatory processes for private rent-seeking. See John T. De-lacourt & Todd J. Zywicki, The FTC and State Action: Evolving Views on the Proper Role ofGovernment, 72 ANTITRUST L.J. 1075 (2005).


trust liability may find it easier to prevent managers from harmingcompetition through exclusion than from doing so through collusion.241

Second, the agencies may bring more collusion cases because they havechosen to direct the bulk of their investigative resources toward collusionrather than exclusion for reasons of efficiency in budget and personnel alloca-tion rather than because of their perception of the distribution of error costs.They likely find it cost-effective to focus their efforts outside merger reviewon anticompetitive conduct that lacks a plausible efficiency justification,242

and when they do, they likely discover that naked collusion is more commonthan plain exclusion.243

Third, the relative counts of collusion and exclusion cases may overstatethe relative frequency of collusion because of the way the cases are counted.The most plausible way this could happen is if mixed cases—cases in whichfirms both collude and exclude—are not infrequent and are routinely viewedsolely as collusion cases.244 Under such circumstances, the agencies may ap-pear to be directing their enforcement efforts toward collusion more than theyare doing in fact.

These three explanations—greater deterrence of exclusion, agency resourceallocation decisions, and treating mixed cases solely as collusion matters—arecollectively more likely than the fourth possibility, which is implicitly ac-cepted by those who favor the error cost argument for downplaying exclusion:that collusion cases are more common because it is more difficult to distin-guish harmful from beneficial conduct in exclusion settings than in collusionsettings. Yet only the fourth explanation tends to suggest that enforcers andcourts are more likely to make errors in resolving exclusion allegations thancollusion allegations.

Moreover, the fourth explanation is problematic. It is hard to see why thedifficulties in identifying harm to competition would be systematically greaterin an exclusion setting than a collusion one if the challenged conduct has no

241 See William H. Page, Optimal Antitrust Remedies: A Synthesis 15–16 (Working Paper2012), available at http://ssrn.com/abstract=2061791.

242 See Creighton et al., supra note 81, at 995 (recommending that enforcers target cheap exclu-sion); First, supra note 80, at 160–62 (recommending new remedial tools for attacking monopo-lization through conduct lacking an efficiency justification).

243 See Jacobson, supra note 80, at 361 (noting that in the context of vertical exclusive dealing,exclusionary conduct lacking any plausible justification “appears unusually rare”). Even if plainexclusion is no less common than naked collusion, moreover, antitrust cases attacking plainexclusion may be less frequent if exclusionary conduct lacking any plausible business justifica-tion is commonly attacked under non-antitrust statutes. See A. Douglas Melamed, ExclusionaryConduct Under the Antitrust Laws: Balancing, Sacrifice, and Refusals to Deal, 20 BERKELEY

TECH. L.J. 1247, 1249 (2005).244 See supra notes 43–48, 149–155 and accompanying text.


plausible efficiency justification in each case—that is, in a universe limited tonaked collusion and plain exclusion. It is also hard to see why the difficultiesidentifying anticompetitive practices would be greater in an exclusion settingfor the rest of the relevant universe—that is, if the challenged conduct doeshave a plausible efficiency justification in each case.

For example, to pick an exclusionary practice,245 a manufacturer’s require-ment that a retailer not distribute rival manufacturers’ products may benefitthe manufacturer primarily because the practice creates efficiencies (as byeliminating double marginalization or otherwise aligning incentives betweenmanufacturer and retailer), or it may benefit the manufacturer primarily be-cause the practice confers market power by excluding rival manufacturersfrom access to low cost distribution. It may be difficult to distinguish betweenthese explanations. But it may equally be difficult to tell whether a retailingjoint venture between two manufacturers, to pick a collusive practice, prima-rily benefits them by lowering production costs (as by generating scale econo-mies) or by conferring market power through a reduction in the directcompetition between them. It may also be difficult to tell whether an agree-ment among rivals to exchange information (perhaps implemented throughtheir trade association) benefits competition by helping the firms match pro-duction to costs and demand, or whether it harms competition by facilitatingcollusion, as by helping them detect cheating rapidly.246

The casual but erroneous supposition that it is harder to distinguish harmsfrom efficiencies in the exclusion setting is likely the result of a categorymistake: comparing the difficulty of proving harm from naked cartels, wherethere is no plausible efficiency, with the difficulty identifying anticompetitiveexclusionary practices when the conduct has a plausible justification.247 Nakedcartels probably come first to mind as a collusive practice, given the fre-quency with which the enforcement agencies announce such cases. But when

245 No example can be representative, but the argument plausibly generalizes beyond the exam-ples provided here.

246 For case examples that suggest the difficulty of distinguishing harms to competition andefficiencies when collusion is alleged, see, for example, Broadcast Music, Inc. v. CBS Inc., 441U.S. 1 (1979) (reversing lower court decision finding a blanket licensing agreement among rivalcopyright owners illegal per se); United States v. Topco Associates, 405 U.S. 596 (1972) (strik-ing territorial restrictions on the marketing of private label products distributed by a joint ventureamong supermarkets, some of which were effectively horizontal rivals); General Motors Corp.,103 F.T.C. 374 (1984) (approving joint production venture between rival automakers subject toconditions).

247 Similarly, it would be inappropriate to make inferences about the relative harm arising fromexclusion and collusion by comparing the competitive harm from the typical naked cartel to thecompetitive harm from the typical instance of exclusionary conduct, whether plain or justified byan efficiency. See Aaron Edlin, Predatory Pricing, in RESEARCH HANDBOOK ON THE ECONOMICS

OF ANTITRUST LAW 144, 173 (Einer Elhauge ed., 2012) (“Presumably, it is true . . . that mostprice cuts are pro-competitive . . . . However, no antitrust proposals attack all price cuts, so thatsample is irrelevant.”).


thinking about exclusion without reflecting on how to make an apples-to-ap-ples comparison, the top-of-mind exclusionary practice will often be a verticalcontract with a plausible efficiency justification, rather than an example ofplain exclusion, such as the conduct found to violate the antitrust laws inLorain Journal and Microsoft. The casual supposition misleads because plainexclusion would correspond to a naked cartel, while a vertical contract with aplausible justification would not. A comparison based on a category mistakeprovides no reason to expect more frequent enforcement and adjudicative er-rors in resolving exclusion cases than in evaluating collusion cases.

B. RELATIVE COST OF ERRORS

For the reasons set forth above, downplaying exclusion cannot be justifiedbased on the view that false positives are more common in the exclusion set-ting. If a justification remains, it would instead have to be based on a supposi-tion that any errors that do occur are more costly when exclusion is allegedthan when collusion is alleged. Two primary arguments have been offered forthis latter supposition—one based on empirical studies and the other rooted inan analysis of institutional competence—but neither is convincing.

First, some commentators suggest that the many empirical studies that haveidentified efficiencies and other competitive benefits from vertical integrationand vertical agreements show that anticompetitive consequences of such prac-tices are unlikely, so antitrust rules should favor defendants in exclusionaryconduct categories.248 But these studies would be probative only if they showthat errors are more costly when exclusion is alleged than when collusion isalleged, which they do not.249 Many of the cited studies do not discriminatebetween exclusionary and collusive explanations for vertical agreements;250

taken at face value, they would question the prevalence of both explanationsand thus cannot provide a basis for downplaying exclusion relative to collu-sion.251 Moreover, the business decisions evaluated in these studies are com-

248 Jeffrey Church, Vertical Mergers, in 2 ISSUES IN COMPETITION LAW AND POLICY, supranote 52, at 1455, 1495–97; James C. Cooper, Luke M. Froeb, Dan O’Brien & Michael G. Vita,Vertical Antitrust Policy as a Problem of Inference, 23 INT’L J. INDUS. ORG. 639 (2005); Geof-frey A. Manne & Joshua D. Wright, Innovation and the Limits of Antitrust, 6 J. COMPETITION L.& ECON. 153, 161 n.23 (2010).

249 The studies do not bear on the relative frequency of errors, so also do not show that falsepositives would be more frequent in exclusion cases than collusion cases.

250 Cf. Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551 U.S. 877, 892–94 (2007) (dis-cussing both collusive and exclusionary explanations for resale price maintenance).

251 The prevention or elimination of free riding can potentially justify both vertical and hori-zontal agreements. Cont’l T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977) (vertical non-priceagreement); Polk Bros., Inc. v. Forest City Enters., Inc., 776 F.2d 185 (7th Cir. 1985) (horizontalmarket division agreement between potential rivals). “Free riding” refers to the externality thatarises when investments by one firm increase demand or reduce costs for rivals, and the first firmis not compensated for providing this benefit. The elimination of free riding is frequently in-


monly made under the shadow of the antitrust laws. Because of the deterrenteffect of antitrust enforcement, the observed practices would be expected dis-proportionately to benefit competition even if they have anticompetitive po-tential in other settings.252 In consequence, empirical studies evaluatingexclusionary conduct provide little evidence of value regarding either the po-tential for those practices to harm competition253 or the likelihood that theparticular instances selected for enforcement in fact harm competition.254

The other commonly offered justification for the view that errors are morecostly when exclusion is alleged than when collusion is alleged turns on aclaim about the institutional competence of enforcers and courts. For institu-tional competence to matter, enforcers and courts must make systematically

voked to justify restrictions imposed by manufacturers on distributors, where the manufacturerclaims that absent the restrictions, the dealer would not provide an appropriate level of servicesto customers or promotional investments. See generally Benjamin Klein & Andres V. Lerner,The Expanded Economics of Free-Riding: How Exclusive Dealing Prevents Free-Riding andCreates Undivided Loyalty, 74 ANTITRUST L.J. 473 (2007); Benjamin Klein & Kevin M. Mur-phy, Vertical Restraints as Contract Enforcement Mechanisms, 31 J.L. & ECON. 265 (1988);Lester G. Telser, Why Should Manufacturers Want Fair Trade? 3 J.L. & ECON. 86 (1960). Forother examples of business justification defenses considered in antitrust cases, see HOVENKAMP,supra note 36, § 5.2 (horizontal joint ventures); Mary Anne Mason & Janet L. McDavid, Busi-ness Justification Defenses, in 2 ISSUES IN COMPETITION LAW AND POLICY supra note 52, at1019 (monopolization cases); Michael A. Salinger, Business Justification Defenses in TyingCases, in 3 ISSUES IN COMPETITION LAW AND POLICY supra, at 1911 (tying cases); Joseph Kat-tan, Efficiencies and Merger Analysis, 62 ANTITRUST L.J. 513 (1994) (horizontal mergers).

252 See Wickelgren, supra note 77, at 55 (“[H]ow often one should expect to see an anticompe-titive manifestation of a restraint will depend on how that restraint is likely to be judged.”); id. at56 (“When considering whether to change the treatment of RPM, for example, from the per serule to the rule of reason, it is not important what fraction of existing uses of RPM are pro- oranticompetitive. Rather, what matters is how many more pro- and anticompetitive instances ofRPM will arise under some version of rule-of-reason treatment rather than per se treatment.”).Relatedly, the leading studies of vertical restraints may have examined competitive effects pri-marily in relatively competitive markets, where those practices would not be expected to harmcompetition, rather than in sectors in which firms exercise substantial market power, where anti-trust enforcement tends to be concentrated. Vincent Verouden, Vertical Agreements: Motivationand Impact, in 2 ISSUES IN COMPETITION LAW AND POLICY, supra note 52, at 1813, 1837. Al-though defendants commonly prevail in vertical restraint cases, those outcomes frequently resultfrom lack of proof of market power so they provide little guide to the likelihood that such con-duct, whether exclusionary or collusive, would be justified when defendants have market power.Furthermore, the prevalence of a practice in markets thought to perform competitively at bestestablishes that the practice could be procompetitive. It does not indicate whether the conductcould harm competition when employed by firms with market power or whether anticompetitiveuses have been deterred by the threat of antitrust enforcement. See Randal D. Heeb et al., supranote 153, at 229 (loyalty rebates, which in theory can under some circumstances benefit competi-tion and under other circumstances harm competition, are sometimes used by cartels).

253 Empirical economic studies about the competitive effects of specific business practices aregenerally more useful for evaluating conduct in industries similar to those studied than for gener-alizing across industries to formulate legal rules. Jonathan B. Baker & Timothy F. Bresnahan,Economic Evidence in Antitrust: Defining Markets and Measuring Market Power, in HANDBOOK

OF ANTITRUST ECONOMICS 1, 24–29 (Paolo Buccirossi ed., 2008).254 By contrast, the many examples of anticompetitive conduct observed during periods of lax

antitrust enforcement suggest the benefits of antitrust. See Baker, supra note 228, at 36–38.


different (and worse) errors when evaluating exclusion cases than when ana-lyzing collusion cases. Because the main concern of those arguing in favor ofdownplaying exclusion is that false positives will chill procompetitive con-duct, the issue turns on the relative incidence and significance of judicial er-rors when the conduct under review has a plausible efficiency justification. Inthe exclusion context, this includes price cutting and new product introduc-tions—conduct that at least in the short run benefits consumers.255 There is noreason to think that enforcers and courts will systematically fail to noticewhen defendants have a plausible efficiency justification in exclusion casesyet recognize that possibility in collusion cases.256 If outcomes are systemati-cally biased in favor of plaintiffs in exclusion cases but not in collusion cases,that outcome would instead have to result from some aspect of the decision-making process that differs across the two settings.

The most common institutional competence argument presumes that exclu-sion cases are disproportionately prompted by the trumped up complaints ofinefficient rivals, losers in the marketplace, that seek to overturn the market’sverdict in the courts directly as plaintiffs or indirectly by inducing enforce-ment agency suits. If so, and if, in addition, complaining rivals bringing badcases tend to have more influence over the judicial process than the firmswrongly accused of anticompetitive exclusionary conduct, then false positiveswould be more likely to arise in exclusion cases than in collusion cases.257

255 Similarly, collusive conduct can appear to benefit consumers in the short run. Examplesmay include various practices facilitating coordination, such as the parallel adoption of simpli-fied and common product definitions, the parallel adoption of price lists, or the parallel adoptionof guarantees to buyers that they will get the best price the seller gives any buyer.

256 To similar effect, some claim that antitrust enforcement against exclusion is problematicbecause it is difficult for courts to make the detailed factual assessments required to determinewhether firms can solve their exclusion problems or to compare the harms from exclusionaryconduct against the procompetitive benefits. See, e.g., Manne & Wright, supra note 248, at 157(characterizing Easterbrook’s analysis as premised in part on the view that “errors of both typesare inevitable, because distinguishing procompetitive conduct from anticompetitive conduct is aninherently difficult task in the single firm context”). Yet if fact finding is the problem withcomprehensive reasonableness review of exclusion allegations, it raises a comparable difficultyfor collusion enforcement under the comprehensive rule of reason, where a court must determinewhether firms can solve their collusion problems and analyze whether the benefits to competitiondissipate or eliminate the harms. See Leegin Creative Leather Prods., Inc. v. PSKS, Inc., 551U.S. 877, 916–17 (2007) (Breyer, J., dissenting) (noting the difficulties of assessing whether thebenefits of resale price maintenance in preventing free riding outweigh the potential harm offacilitating a dealer cartel, and the difficulties judges and juries may face in evaluating marketpower).

257 See, e.g., WILLIAM J. BAUMOL, ROBERT E. LITAN & CARL J. SCHRAMM, GOOD CAPITALISM,BAD CAPITALISM, AND THE ECONOMICS OF GROWTH AND PROSPERITY 118–19 (2007); Edward A.Snyder & Thomas E. Kauper, Misuse of the Antitrust Laws: The Competitor Plaintiff, 90 MICH.L. REV. 551 (1991) (noting that a small minority of the 37 horizontal restraints cases filed bycompetitor plaintiffs between 1973 and 1983 in five federal districts alleging exclusionary prac-tices seemed meritorious). If the courts do not weed out false claims from competitors, moreover,even efficient rivals would be expected to bring unwarranted exclusion claims in order to dis-courage hard competition from dominant firms.


Under such circumstances, antitrust institutions would inappropriately tend toprotect competitors rather than competition in exclusion cases, but not sooften in collusion cases, consistent with what those making this institutionalcompetence argument contend.258

This argument about institutional competence is unconvincing, however,because there is no reason to think that the agencies and courts are biased infavor of the victims of alleged exclusion259 or that unsuccessful rivals cansystematically convince the enforcement agencies and courts to accept badcases.260 Even if unsuccessful rivals or terminated dealers foresee the possibil-ity of substantial gains from bringing a speculative (or even trumped up) anti-trust complaint, they must also consider their low probability of success inevaluating their expected gain, and thus in deciding whether to bring a case.After all, it is no more difficult for enforcers and courts to understand thepossible biases of rivals, and discount their testimony appropriately, than forthose decision makers to discount as necessary the testimony of alleged ex-cluding firms and customers.261 More than most firms, moreover, defendantsin exclusion cases, particularly large firms accused of monopolization, tend tohave the ability to present an effective courtroom case, employing top qualitylegal representation and economic experts and supporting them with a gener-ous budget.262 Large firm defendants in exclusion cases also tend to have theresources to make an effective public relations case and mobilize politicalsupport.263

258 See Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 488 (1977) (quotingBrown Shoe Co. v. United States, 370 U.S. 294, 320 (1962) (noting that antitrust aims to protect“competition, not competitors”)).

259 One commentator speculates, without evidence, that antitrust enforcers tend to sympathizewith smaller firms. D. Daniel Sokol, The Strategic Use of Public and Private Litigation in Anti-trust as Business Strategy, 85 S. CAL. L. REV. 689, 730–31 (2012). It is also possible that juriescould systematically misinterpret colorful evidence of defendant intent to crush rivals as indicat-ing an aim to do so through anticompetitive means (rather than by lowering costs and prices orintroducing new or better products), notwithstanding jury instructions making the relevant dis-tinction. The likelihood and magnitude of the possible prejudicial effect of such evidence on theinterpretation of aggressively competitive conduct close to the line is hard to assess, but if thispossibility is important systematically, it is better addressed through rulings on the admissibilityof evidence in those cases where the problem may arise rather than through caution in enforcingthe antitrust laws against anticompetitive exclusion generally.

260 In addition, the antitrust injury requirement limits the possible misuse of the antitrust lawsin this way.

261 See FCC, STAFF ANALYSIS AND FINDINGS, WT Docket No. 11-65, 44–45 & n.255 (Nov. 29,2011), available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DA-11-1955A2.pdf(describing interests of merging firms and merger opponents and their possible alignment withthe public interest).

262 But see Sokol, supra note 259, at 731 (arguing that dominant firms do not employ effectivecounterstrategies because they are prone to inertia, focused on running their business, arrogant,and likely to be viewed as unsympathetic victims).

263 Another institutional competence argument concerns remedy. Exclusionary violations aresometimes said to be difficult to remedy in rapidly changing industries, where dramatic changes


C. OTHER ERROR COST ARGUMENTS

Other error cost arguments for giving priority to collusion over exclusionare also unconvincing. Some suggest, consistent with Justice Scalia’s Trinkodicta,264 that false negatives are limited in antitrust cases because markets arealmost invariably self-correcting265 or that false positives are particularly ex-pensive to society because market power rather than competition forms theprimary spur to innovation.266 Those controversial claims should not be ac-cepted.267 Market power is often durable: economic theory suggests many rea-sons why monopoly power would not be transitory,268 and the case law offersmany examples of durable market power,269 including in high-tech markets.270

Moreover, the empirical evidence indicates that the push of competition isgenerally more important for innovation than the pull of monopoly.271 Hence afocus on “dynamic competition” does not justify exclusionary conduct likemonopolization.272 For the present discussion, however, the more important

in the marketplace are likely to occur between the date of violation and the time a court deter-mines liability and crafts relief. Even when remedies that would restore competition in the mar-ket under review appear limited, however, remedies providing general deterrence (such as finesand damages) remain available. See United States v. Microsoft Corp., 253 F.3d 34, 49 (D.C. Cir.2001); Louis Kaplow, An Economic Approach to Price Fixing, 77 ANTITRUST L.J. 343, 416–32(2011) (preferring fines to injunctions as the sanction for collusion on general deterrencegrounds).

264 See supra notes 12–20 and accompanying text.265 See, e.g., Fred S. McChesney, Easterbrook on Errors, 6 J. COMPETITION L. & ECON. 11, 16

(2010); Fred S. McChesney, Talking ‘Bout My Antitrust Generation: Competition for and in theField of Competition Law, 52 EMORY L.J. 1401, 1412 (2003); see also Hylton, supra note 225, at102; Manne & Wright, supra note 248, at 157 (claiming that Easterbrook’s analysis is premisedin part on the view that “false positives are more costly than false negatives, because self-correc-tion mechanisms mitigate the latter but not the former”).

266 See Evans & Hylton, supra note 166, at 203.267 For a discussion of other arguments potentially related to the balance of error costs in the

context of monopolization enforcement, see Baker, supra note 21, at 616–20.268 See, e.g., Ariel Ezrachi & David Gilo, Are Excessive Prices Really Self-Correcting? 5 J.

COMPETITION L. & ECON. 249 (2008) (noting that supracompetitive prices only attract entryefforts if they signal that the post-entry price would be high or that the incumbent firms havehigh costs, and even then entry may not succeed in competing those prices down to competitivelevels); see also Oliver E. Williamson, Delimiting Antitrust, 76 GEO. L.J. 271, 289 (1987)(“Economies as an antitrust defense excepted, no one has provided a demonstration that the costdifferences are as Easterbrook indicates. Easterbrook has an undischarged burden of proof thatthe cost of false positives in the market power region where strategic behavior is implicated issimilarly low.”).

269 See, e.g., Standard Oil Co. v. United States, 221 U.S. 1 (1911); United States v. DentsplyInt’l, Inc., 399 F.3d 181 (3d Cir. 2005).

270 See, e.g., United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001).271 See generally Baker, supra note 141; Shapiro, supra note 141.272 Baker, supra note 166; cf. Catalano, Inc. v. Target Sales, Inc., 446 U.S. 643, 649 (1980)

(rejecting the argument that the potential for supracompetitive prices to induce entry could justifyhorizontal price fixing). Moreover, the suggestion that an increased financial reward to defend-ants promotes competition by increasing their incentive to invest and innovate, if offered withoutqualification or recognition of tradeoffs, has no logical stopping point. It would imply that the


point is that even if these suspect claims were accepted, they would not justifythe rhetorical consensus prioritizing collusion: they would be reasons to op-pose all antitrust enforcement, not to downgrade exclusion relative tocollusion.

Another error cost argument has been accepted by some U.S. courts as areason to allow a monopolist to make exclusive vertical agreements: the claimthat exclusionary practices cannot make matters worse (and thus cannot harmcompetition) because there is a “single monopoly profit.”273 This possibilitydoes not justify treating exclusion less seriously than collusion, however, be-cause the argument applies only in narrow circumstances.274 If the excludingfirms have literally no fringe rivals and face no potential entrants, and if thereare no ways that buyers can substitute away from the monopoly, thenthere may indeed be no way to increase the rents from exercising marketpower through (further) exclusionary conduct. Outside of such unusual facts,though, firms can potentially obtain, extend, or maintain their market powerthrough exclusionary conduct that suppresses these forms of competition,275

government should subsidize firms heavily, allowing them to invest even more for the benefit ofsociety. See Hylton & Lin, supra note 167, at 258–59 (noting that exclusionary practices canencourage investment by making it more profitable).

273 E & L Consulting, Ltd. v. Doman Indus. Ltd., 472 F.3d 23 (2d Cir. 2006); G.K.A. BeverageCorp. v. Honickman, 55 F.3d 762, 767 (2d Cir. 1995); Town of Concord v. Boston Edison Co.,915 F.2d 17, 23, 32 (1st Cir. 1990) (Breyer, C.J.); see Jefferson Parish Hosp. Dist. No. 2 v. Hyde,466 U.S. 2, 36–37 (1984) (O’Connor, J., concurring).

274 Similarly, a horizontal agreement with no efficiency justification would not harm competi-tion if the horizontal rivals are already coordinating perfectly. But this unlikely possibility doesnot justify downplaying the concern with collusive conduct.

275 See generally ANTITRUST LAW IN PERSPECTIVE, supra note 25, at 417–18 (example inwhich a monopolist manufacturer harms competition by consolidating distribution in one dealer);id. at 861–65 (example in which single monopoly profit theory holds when downstream buyeruses monopolized product in fixed proportions with other inputs but fails to hold when the prod-uct is used in flexible proportions); id. at 811–12 (example in which a monopolist achievesadditional market power through the exclusionary effect of tying); see also Timothy F. Bresna-han, Monopolization and the Fading Dominant Firm, in COMPETITION LAW AND ECONOMICS:ADVANCES IN COMPETITION POLICY ENFORCEMENT IN THE EU AND NORTH AMERICA 264 (AbelM. Mateus & Teresa Moreira eds., 2010) (demonstrating that a dominant firm threatened by rivalinnovation can profit by blocking those rivals, leading to the failure of the single monopoly profittheory in the case of technologically dynamic industries). An incumbent monopolist would alsobe unable to increase its market power through exclusion in an unusual case in which it hassufficient bargaining power to permit efficient entry while appropriating virtually all the rents.See Phillipe Aghion & Patrick Bolton, Contracts as a Barrier to Entry, 77 AM. ECON. REV. 388(1987) (discussing a model in which the manufacturer and distributor seek to allow efficiententry and extract all the rents the entrant creates, but bargaining over splitting the surplus canbreak down when the parties have imperfect information). The economics literature has alsoconsidered the applicability of the “single monopoly profit” argument in the context of “monop-oly leveraging” concerns outside the scope of the present discussion. See generally Patrick Rey& Jean Tirole, A Primer on Foreclosure, in 3 HANDBOOK OF INDUSTRIAL ORGANIZATION 2145,2182–83 (Mark Armstrong & Robert Porter eds., 2007) (discussing models of “horizontal fore-closure”); Salop & Romaine, supra note 218, at 623–26 (distinguishing between application of


even if the excluded firms are less efficient competitors than the excludingfirms.276

In sum, the policy (or error cost) arguments for downplaying exclusion donot stand up to analysis, whether they are grounded in common ideas aboutthe difficulty of distinguishing procompetitive conduct from anticompetitiveexclusion or in the more controversial arguments made in Trinko. The closerelationship between the ways by which exclusion and collusion allow firmsto exercise market power, and the convergence in the legal rules governingexclusionary and collusive conduct, do not mislead. Exclusion should betreated as seriously as collusion.

VI. IMPLICATIONS FOR ENFORCEMENT

Exclusion should be recognized as a core concern of competition law andpolicy along with collusion, and the use of the common rhetorical conventionthat treats anticompetitive exclusionary conduct as of lesser importance thananticompetitive collusion should be avoided. Doing so could lead enforcers toplace a higher priority on challenging exclusion than they do today, particu-larly aiming to prevent exclusionary conduct that forecloses potential entry inmarkets subject to rapid technological change. It is particularly important toreaffirm the innovation benefits of antitrust enforcement against anticompeti-tive exclusion in high-tech markets in the wake of the Trinko opinion’s nodtoward monopoly power as a means of encouraging innovation,277 which risksleading lower courts astray.278

Recognizing exclusion as a core competition problem is unlikely to leadcourts to modify the substantive antitrust rules they employ to test exclusion-ary conduct. Those rules are, in general, well-designed to test the reasonable-ness of firm conduct, whether the analysis is truncated or comprehensive. Noris rhetorical parity likely to affect the frequency with which the courts addressexclusionary conduct. Because the rules would not change, it is unlikely thatprivate plaintiffs, which account for the bulk of antitrust litigation, wouldbring more exclusion cases (other than follow-ons in the event governmentactions against exclusionary conduct increase).279 Government enforcers

the “single monopoly profit” argument to monopoly leveraging allegations and preserving mo-nopoly allegations).

276 See supra note 189.277 See supra note 16 and accompanying text.278 See supra notes 167–172 and accompanying text.279 Neither the relative frequency of the underlying anticompetitive conduct nor the tools avail-

able to enforcers and plaintiffs for identifying them would directly be affected if exclusion is nolonger described as a lesser antitrust offense. The number of exclusion cases could even decline.To the extent firms today have been misled by the common rhetoric, and incorrectly believe thatanticompetitive exclusionary conduct would not successfully be challenged, a rhetorical change


should treat exclusionary conduct as comparable in priority with collusiveconduct, although, as a practical matter, the relative frequency of governmentcases alleging anticompetitive exclusion would increase only to the extent theenforcement agencies shift resources away from investigating and challenginganticompetitive collusion.280 Even if the agencies targeted exclusionary con-duct aggressively, budgetary and staffing limitations would most likely permitthe agencies to bring only a handful of additional exclusion cases annually.281

The major benefit of recognizing that exclusion is as important as collusionwould instead come from protecting the legitimacy of the antitrust rules gov-erning exclusionary conduct against pressure for modifications that wouldlimit enforcement inappropriately. Enforcers and courts would not be misledby the contemporary consensus in antitrust discourse that shies away fromattacking anticompetitive exclusion, instead urging a focus on collusive con-duct. A rhetorical shift may also heighten the salience of addressing the openquestions in the formulation of the rules governing truncated condemnation ofanticompetitive exclusion,282 and thereby encourage the further developmentof the law in that area.

A shift in how exclusion is viewed could also matter for remedies, as it mayencourage the Justice Department to raise the penalties for anticompetitiveexclusionary conduct, in appropriate cases, through criminal enforcement.283

The Justice Department has the discretion to challenge anticompetitive exclu-sionary conduct as a civil violation or to prosecute it criminally, in the sameway that the government has the discretion to attack collusion civilly or crimi-

of course could increase deterrence, reduce the prevalence of such conduct, and, in consequence,reduce the frequency with which it is challenged.

280 The frequency of government enforcement in various categories is surveyed in Kovacic,supra note 30, at 407–76.

281 In recent years, the Justice Department has brought between two and five non-merger civilactions annually and challenged twelve to twenty mergers during each of the least five years(fiscal years 2007–2011), ANTITRUST DIV., U.S. DEP’T OF JUSTICE, WORKLOAD STATISTICS FY2002–2011, http://www.justice.gov/atr/public/workload-statistics.html, and so has limited abilityto increase the number of exclusion cases through reallocation of civil resources. The AntitrustDivision also filed between forty and ninety criminal cases annually during these years. Becausecriminal exclusionary conduct cases are likely to be rare even if exclusion is viewed as havingequal priority as collusion, the Justice Department is unlikely to be able to increase its exclusion-ary conduct case count substantially without shifting resources from criminal to civil investiga-tions. The Federal Trade Commission would similarly have only limited ability to increase thenumber of exclusion cases through reallocation of resources.

282 See generally supra notes 117–138 and accompanying text.283 The penalties could also be raised by awarding the government the ability to collect civil

fines for antitrust violations. See First, supra note 80, at 153–65 (recommending that Congressenlarge government antitrust remedies to include civil penalties, and that the enforcement agen-cies initially target the use of those remedies to monopolization cases in which the exclusionaryconduct had no efficiency justification or in which the defendant engaged in a systemic effort tomaintain monopoly).


nally.284 Criminal enforcement has been employed to attack exclusion in thepast, as with a monopolization case brought against a dominant newspaperand its senior officials alleging exclusionary conduct similar to the anticompe-titive practices attacked in Lorain Journal.285 Today, however, criminal anti-trust enforcement is directed at cartel conduct,286 consistent with the commondescription of exclusion as a lesser offense.

If exclusion is viewed as central to antitrust, criminal prosecution would nolonger be reserved for collusive conduct. When applied to exclusionary con-duct, it would almost surely be directed at the most egregious cases of plainexclusion, in much the way that the government now targets only the mostegregious naked cartels for indictment.287 It would be easy to imagine a crimi-nal antitrust indictment (as well as other criminal charges) brought againstsenior executives if, for example, a dominant firm harmed competition by

284 The Justice Department challenged the lysine cartel criminally, for example, but brought acivil case when challenging price fixing among the major airlines. Compare United States v.Andreas, 216 F.3d 645, 680 (7th Cir. 2000) (upholding criminal convictions of executives con-spiring to fix lysine prices), with United States v. Airline Tariff Publ’g Co., No. 92–2854 (SSH),1994 WL 502091 (D.D.C. Aug. 10, 1994) (final consent decree settling airline price-fixingallegations).

285 Kansas City Star v. United States, 240 F.2d 643 (1957); see also United States v. HiltonHotels Corp., 467 F.2d 1000 (9th Cir. 1972) (upholding criminal conviction of firms that con-spiring to boycott suppliers, though the group boycott was collusive rather than exclusionary);William E. Kovacic, The Intellectual DNA of Modern U.S. Competition Law for Dominant FirmConduct: The Chicago/Harvard Double Helix, 2007 COLUM. BUS. L. REV. 1, 17–18 & n.44(identifying three criminal monopolization cases brought during the early 1960s). Only a smallfraction of antitrust cases prosecuted criminally have involved exclusionary practices. See JosephC. Gallo et al., Criminal Penalties Under the Sherman Act: A Study of Law and Economics, 16RES. IN L. & ECON. 25, 28 (Richard O. Zerbe, Jr. ed., 1994) (finding that 33, or 2 percent, of the1,522 criminal antitrust cases brought by the Justice Department between 1955 and 1993 in-volved exclusionary practices).

286 “In general, current [Antitrust] Division policy is to proceed by criminal investigation andprosecution in cases involving horizontal, per se unlawful agreements such as price fixing, bidrigging, and customer and territorial allocations. . . . [C]ivil prosecution is used with respect toother suspected antitrust violations . . . .” U.S. DEP’T. OF JUSTICE, ANTITRUST DIVISION MANUAL

III-12 (5th ed. Nov. 2010), available at http://www.justice.gov/atr/public/divisionmanual/atrdivman.pdf.; accord, Thomas O. Barnett, Assistant Att’y Gen., Antitrust Div., U.S. Dep’t ofJustice, Criminal Enforcement of Antitrust Laws: The U.S. Model (Sept. 14, 2006), available athttp://www.justice.gov/atr/public/speeches/218336.htm (“[T]he Division focuses its criminal en-forcement . . . . narrowly on price fixing, bid-rigging, and market allocations, as opposed to the‘rule of reason’ or monopolization analyses used in civil antitrust law.”). In the past, however,the Antitrust Division has indicated that criminal enforcement is appropriate when predatory(exclusionary) conduct supports collusion. See REPORT OF THE ATTORNEY GENERAL’S NATIONAL

COMMITTEE TO STUDY THE ANTITRUST LAWS 350 (1955) (statement of Assistant AttorneyGeneral).

287 See Kovacic, supra note 30, at 416–23 (describing the progressive evolution of the U.S.norm treating cartel behavior as criminal conduct since the 1970s, and providing statistics con-cerning the relative frequency of criminal (DOJ) and civil (FTC) enforcement against anticompe-titive horizontal agreements). See generally Donald I. Baker, To Indict or Not to Indict:Prosecutorial Discretion in Sherman Act Enforcement, 63 CORNELL L. REV. 405 (1978).


destroying its key rival’s factory,288 or if a price-fixing cartel engaged in coop-erative conduct to exclude an actual or potential rival that threatened todestabilize its collusive arrangement.289 Criminal enforcement in exclusioncases would be rare—cartel cases would surely remain the mainstay of theJustice Department’s criminal enforcement efforts—but that is not a reason toavoid criminal sanctions in appropriate exclusionary conduct cases.

VII. CONCLUSION

Enforcers and commentators routinely describe anticompetitive exclusionas a lesser offense than anticompetitive collusion. The absence of rhetoricalparity misleads because the two types of conduct harm competition in similarways and are treated comparably in the framing of antitrust rules. Nor dopolicy considerations, whether or not discussed in “error cost” terms, suggestdownplaying exclusion relative to collusion in antitrust enforcement.

The rhetorical relegation of anticompetitive exclusion to antitrust’s periph-ery must end. The more that exclusion is described as a lesser offense, themore its legitimacy as a subject for antitrust enforcement will be underminedand the greater the likelihood that antitrust rules will eventually change tolimit enforcement against anticompetitive foreclosure when they should not. Itis time to recognize that exclusion, like collusion, is at the core of soundcompetition policy.

288 See supra note 55 (providing examples of egregious exclusionary conduct, including de-struction of a rival’s in-store displays, espionage and sabotage, false statements disparaging arival’s product, and destruction of a rival’s retail store); see also United States v. Empire GasCorp., 537 F.2d 296, 298 & n.1 (8th Cir. 1976) (involving a corporate president acquitted incriminal case alleging destruction of property; the Justice Department also brought an unsuccess-ful attempt to monopolize case against the firm).

289 The moral condemnation and loss of liberty associated with criminal sanctions would thusbe limited to the perpetrators of serious anticompetitive conduct that firms and their managersshould have understood in advance would be subject to criminal prosecution.

exclusion as a core competition concern - digital commons

Documents